{ "meta": { "title": "Awesome Game & Embodied Agents", "snapshotDate": "2026-09-14", "paperWindow": { "from": "2021-09-14", "to": "2026-09-14" }, "arxivWindow": { "from": "2025-09-14", "to": "2026-09-14" }, "sourceSnapshot": { "rows": 106, "emptyRows": 19, "duplicateRows": 4, "uniqueTitledRows": 83 }, "knowledgeUrl": "https://dy8q0bnq8y.feishu.cn/wiki/KAI9wbLEniLIIgkHAi6cU48bndd", "referenceUrl": "https://github.com/kairunwen/Awesome-Robot-Use-Agent#projects", "counts": { "articles": 6, "papers": 83, "projects": 5, "datasets": 4, "benchmarks": 8, "demos": 17 }, "repositoryUrl": "https://github.com/LAMDA-NeSy/Awesome-Game-Embodied-Agents", "siteUrl": "https://lamda-nesy.github.io/Awesome-Game-Embodied-Agents/", "ratingsApiUrl": "https://awesome-gea-ratings.ratings-worker.workers.dev/ratings", "language": "en" }, "resources": [ { "id": "kb-g05", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "DreamerV3", "title": "Mastering diverse control tasks through world models", "domains": [ "game", "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "Nature 2025", "summary": "Trains policies inside a learned world model with a shared algorithm configuration across diverse control tasks, connecting game and robot learning.", "tags": [ "World models", "Reinforcement learning" ], "environment": "DMLab30, including four Lasertag tasks; also Atari, Minecraft, and other environments.", "interface": "Task-specific observations, actions, and rewards feed an RSSM world model; actor–critic learning uses imagined trajectories. Each task is trained separately.", "limits": "Fixed hyperparameters do not imply a single policy with shared weights. The DMLab aggregate is not a Lasertag score, and different interaction budgets prevent direct ranking. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://www.nature.com/articles/s41586-025-08744-2", "code": "https://github.com/danijar/dreamerv3", "pdf": "https://www.nature.com/articles/s41586-025-08744-2.pdf" }, "rank": 1, "sourceId": "G05", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS", "Multi-game" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.nature.com/articles/s41586-025-08744-2", "checkedAt": "2026-09-14", "authors": [ "Danijar Hafner", "Jurgis Pasukonis", "Jimmy Ba", "Timothy Lillicrap" ], "date": "2025-04-02", "dateLabel": "Publisher record date", "doi": "10.1038/s41586-025-08744-2", "bibtex": "@misc{kb-g05,\n title = {Mastering diverse control tasks through world models},\n author = {Danijar Hafner and Jurgis Pasukonis and Jimmy Ba and Timothy Lillicrap},\n year = {2025},\n url = {https://www.nature.com/articles/s41586-025-08744-2},\n doi = {10.1038/s41586-025-08744-2}\n}", "preview": "assets/papers/kb-g05.png", "previewSourceUrl": "https://www.nature.com/articles/s41586-025-08744-2.pdf" }, "metrics": { "citations": { "value": 172, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4409147637", "updatedAt": "2026-10-19", "recordTitle": "Mastering diverse control tasks through world models" }, "github": { "value": 3893, "source": "GitHub", "sourceUrl": "https://github.com/danijar/dreamerv3/stargazers", "repository": "danijar/dreamerv3", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-f02", "name": "Diffusion for Human Behaviour", "title": "Imitating Human Behaviour with Diffusion Models", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICLR 2023", "summary": "Uses conditional diffusion to jointly predict mouse movement and firing, capturing multiple plausible choices in human behavior.", "tags": [ "Imitation learning", "Diffusion models" ], "environment": "CS:GO Aim Train; also robotic Kitchen and illustrative toy environments.", "interface": "Aim-training images map to joint horizontal/vertical mouse movements and firing, in a half-speed environment.", "limits": "Evaluation uses a fixed-position, half-speed task. Distribution distance and kills measure different outcomes; additional denoising also increases latency. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2301.10677", "code": "https://github.com/microsoft/Imitating-Human-Behaviour-w-Diffusion", "pdf": "https://arxiv.org/pdf/2301.10677" }, "rank": 5, "sourceId": "F02", "sourceEvidence": "Spot checks of key claims", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2301.10677", "checkedAt": "2026-09-14", "authors": [ "Tim Pearce", "Tabish Rashid", "Anssi Kanervisto", "Dave Bignell", "Mingfei Sun", "Raluca Georgescu", "Sergio Valcarcel Macua", "Shan Zheng Tan", "Ida Momennejad", "Katja Hofmann", "Sam Devlin" ], "date": "2023-01-25", "dateLabel": "First arXiv submission", "arxivId": "2301.10677", "bibtex": "@misc{kb-f02,\n title = {Imitating Human Behaviour with Diffusion Models},\n author = {Tim Pearce and Tabish Rashid and Anssi Kanervisto and Dave Bignell and Mingfei Sun and Raluca Georgescu and Sergio Valcarcel Macua and Shan Zheng Tan and Ida Momennejad and Katja Hofmann and Sam Devlin},\n year = {2023},\n url = {https://arxiv.org/abs/2301.10677}\n}", "preview": "assets/papers/kb-f02.png", "previewSourceUrl": "https://arxiv.org/pdf/2301.10677" }, "metrics": { "citations": { "value": 43, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4318242039", "updatedAt": "2026-10-19", "recordTitle": "Imitating Human Behaviour with Diffusion Models" }, "github": { "value": 156, "source": "GitHub", "sourceUrl": "https://github.com/microsoft/Imitating-Human-Behaviour-w-Diffusion/stargazers", "repository": "microsoft/Imitating-Human-Behaviour-w-Diffusion", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-f07", "name": "F.A.C.U.L.", "title": "F.A.C.U.L.: Language-Based Interaction with AI Companions in Gaming", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "AAAI 2026", "summary": "Turns player language into action programs for game companions through spatial grounding, intent routing, and behavior-tree execution.", "tags": [ "Language & action", "Multi-agent learning" ], "environment": "Arena Breakout: Infinite.", "interface": "Player language, game assets, positions, and orientations feed intent routing and object grounding, followed by atomic action programs or behavior trees.", "limits": "This is high-level companion interaction, not frame-level aiming. Throughput gains do not imply equal per-request speedups; earlier project metrics and final paper metrics must be separated. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://ojs.aaai.org/index.php/AAAI/article/view/38842", "code": "https://github.com/tencent-facul/code", "pdf": "https://ojs.aaai.org/index.php/AAAI/article/download/38842/42804" }, "rank": 7, "sourceId": "F07", "sourceEvidence": "Spot checks of key claims", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7137844787", "checkedAt": "2026-09-14", "authors": [ "Wenya Wei", "Sipeng Yang", "Qixian Zhou", "Ruochen Liu", "Xuelei Zhang", "Yifu Yuan", "Yan Jiang", "Yongle Luo", "Hailong Wang", "Tianzhou Wang", "Peipei Jin", "Wangtong Liu", "Zhou Zhao", "Xiaogang Jin", "Elvis S. Liu" ], "date": "2026-03-14", "dateLabel": "Indexed publication date (OpenAlex)", "bibtex": "@misc{kb-f07,\n title = {F.A.C.U.L.: Language-Based Interaction with AI Companions in Gaming},\n author = {Wenya Wei and Sipeng Yang and Qixian Zhou and Ruochen Liu and Xuelei Zhang and Yifu Yuan and Yan Jiang and Yongle Luo and Hailong Wang and Tianzhou Wang and Peipei Jin and Wangtong Liu and Zhou Zhao and Xiaogang Jin and Elvis S. Liu},\n year = {2026},\n url = {https://ojs.aaai.org/index.php/AAAI/article/view/38842}\n}", "preview": "assets/papers/kb-f07.png", "previewSourceUrl": "https://ojs.aaai.org/index.php/AAAI/article/download/38842/42804" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7137844787", "updatedAt": "2026-10-19", "recordTitle": "F.A.C.U.L.: Language-Based Interaction with AI Companions in Gaming" }, "github": { "value": 0, "source": "GitHub", "sourceUrl": "https://github.com/tencent-facul/code/stargazers", "repository": "tencent-facul/code", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-r01", "name": "Principled Video Representations", "title": "Towards Principled Representation Learning from Videos for Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICLR 2024", "summary": "Studies how video pretraining without action labels affects control, including sensitivity to distracting visual dynamics.", "tags": [ "Representation learning", "Reinforcement learning" ], "environment": "ViZDoom Basic and Defend the Center; also GridWorld.", "interface": "Unlabeled video pretraining produces a frozen encoder for PPO control. Action-labeled inverse models provide comparisons.", "limits": "Retain the artificial external distractions, frozen-encoder setting, and theoretical assumptions. Failure of a particular video objective does not establish that all video foundation models are ineffective. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/eb60f103645fd14c94826665f8505a6d-Abstract-Conference.html", "code": "https://github.com/microsoft/Intrepid", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2024/file/eb60f103645fd14c94826665f8505a6d-Paper-Conference.pdf" }, "rank": 23, "sourceId": "R01", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/eb60f103645fd14c94826665f8505a6d-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Dipendra Kumar Misra", "Akanksha Saran", "Tengyang Xie", "Alex Lamb", "John Langford" ], "date": "2024-05-31", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-r01,\n title = {Towards Principled Representation Learning from Videos for Reinforcement Learning},\n author = {Dipendra Kumar Misra and Akanksha Saran and Tengyang Xie and Alex Lamb and John Langford},\n year = {2024},\n url = {https://proceedings.iclr.cc/paper_files/paper/2024/hash/eb60f103645fd14c94826665f8505a6d-Abstract-Conference.html}\n}", "preview": "assets/papers/kb-r01.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/file/eb60f103645fd14c94826665f8505a6d-Paper-Conference.pdf" }, "metrics": { "github": { "value": 36, "source": "GitHub", "sourceUrl": "https://github.com/microsoft/Intrepid/stargazers", "repository": "microsoft/Intrepid", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-r02", "name": "DuRND", "title": "Catching Two Birds with One Stone: Reward Shaping with Dual Random Networks for Balancing Exploration and Exploitation", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "ICML 2025", "summary": "Combines novelty and reward relevance to balance exploration and exploitation in sparse-reward tasks.", "tags": [ "Reinforcement learning", "Exploration & skills" ], "environment": "ViZDoom DefendLine, SaveCenter, SlayGhosts, and CollectKit; also Atari and MiniWorld.", "interface": "ViZDoom observations feed dual random-network auxiliary rewards and PPO. Exact preprocessing and the complete action configuration need further verification.", "limits": "Two networks do not constitute a multi-agent system. Reward correlation is not causal contribution; CollectKit has no firing, and custom returns are not comparable across scenarios. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v267/ma25j.html", "code": "https://github.com/mahaozhe/DuRND", "pdf": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/ma25j/ma25j.pdf" }, "rank": 24, "sourceId": "R02", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v267/ma25j.html", "checkedAt": "2026-09-14", "authors": [ "Haozhe Ma", "Fangling Li", "Jing Yu Lim", "Zhengding Luo", "Thanh Vinh Vo", "Tze-Yun Leong" ], "date": "2025-10-06", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-r02,\n title = {Catching Two Birds with One Stone: Reward Shaping with Dual Random Networks for Balancing Exploration and Exploitation},\n author = {Haozhe Ma and Fangling Li and Jing Yu Lim and Zhengding Luo and Thanh Vinh Vo and Tze-Yun Leong},\n year = {2025},\n url = {https://proceedings.mlr.press/v267/ma25j.html}\n}", "preview": "assets/papers/kb-r02.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/ma25j/ma25j.pdf" }, "metrics": { "github": { "value": 8, "source": "GitHub", "sourceUrl": "https://github.com/mahaozhe/DuRND/stargazers", "repository": "mahaozhe/DuRND", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-r03", "name": "Action-Sufficient Representations", "title": "Action-Sufficient State Representation Learning for Control with Structural Constraints", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "ICML 2022", "summary": "Learns compact state representations that retain control-relevant factors and uses imagined transitions to support policy learning.", "tags": [ "Representation learning", "World models" ], "environment": "ViZDoom Take Cover; also CarRacing.", "interface": "Image, action, and reward sequences produce structured latent states for DQN. Take Cover uses left/right movement only.", "limits": "The Doom metric is survival time, not kills. CarRacing results do not establish FPS performance; action sufficiency depends on structural-model assumptions. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v162/huang22f.html", "pdf": "https://proceedings.mlr.press/v162/huang22f/huang22f.pdf" }, "rank": 25, "sourceId": "R03", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v162/huang22f.html", "checkedAt": "2026-09-14", "authors": [ "Biwei Huang", "Chaochao Lu", "Liu Leqi", "Jose Miguel Hernandez-Lobato", "Clark Glymour", "Bernhard Schölkopf", "Kun Zhang" ], "date": "2022-06-28", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-r03,\n title = {Action-Sufficient State Representation Learning for Control with Structural Constraints},\n author = {Biwei Huang and Chaochao Lu and Liu Leqi and Jose Miguel Hernandez-Lobato and Clark Glymour and Bernhard Schölkopf and Kun Zhang},\n year = {2022},\n url = {https://proceedings.mlr.press/v162/huang22f.html}\n}", "preview": "assets/papers/kb-r03.png", "previewSourceUrl": "https://proceedings.mlr.press/v162/huang22f/huang22f.pdf" }, "metrics": { "citations": { "value": 11, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3207211828", "updatedAt": "2026-10-19", "recordTitle": "Action-Sufficient State Representation Learning for Control with Structural Constraints" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-b01", "name": "COOM", "title": "COOM: A Game Benchmark for Continual Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers", "benchmarks" ], "year": 2023, "venue": "NeurIPS 2023 Datasets and Benchmarks", "summary": "Evaluates continual reinforcement learning under sequential visual and task shifts, tracking forgetting, forward transfer, and new-task learning.", "tags": [ "Continual learning", "Benchmarks" ], "environment": "Doom / ViZDoom.", "interface": "Sequences of visual ViZDoom tasks are evaluated with SAC and continual-learning methods. Task boundaries are known; some methods use task identity.", "limits": "Compare performance, forgetting, and forward transfer together. Known task boundaries or identities differ from an open task stream; normalized average performance is not a win rate. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/d61d9f4fe4357296cb658795fd7999f0-Abstract-Datasets_and_Benchmarks.html", "code": "https://github.com/TTomilin/COOM", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/d61d9f4fe4357296cb658795fd7999f0-Paper-Datasets_and_Benchmarks.pdf" }, "rank": 13, "sourceId": "B01", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/d61d9f4fe4357296cb658795fd7999f0-Abstract-Datasets_and_Benchmarks.html", "checkedAt": "2026-09-14", "authors": [ "Tristan Tomilin", "Meng Fang", "Yudi Zhang", "Mykola Pechenizkiy" ], "date": "2023-12-15", "dateLabel": "Publisher record date", "doi": "10.52202/075280-2965", "bibtex": "@misc{kb-b01,\n title = {COOM: A Game Benchmark for Continual Reinforcement Learning},\n author = {Tristan Tomilin and Meng Fang and Yudi Zhang and Mykola Pechenizkiy},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/d61d9f4fe4357296cb658795fd7999f0-Abstract-Datasets_and_Benchmarks.html},\n doi = {10.52202/075280-2965}\n}", "preview": "assets/papers/kb-b01.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/d61d9f4fe4357296cb658795fd7999f0-Paper-Datasets_and_Benchmarks.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7133239981", "updatedAt": "2026-10-19", "recordTitle": "COOM: A Game Benchmark for Continual Reinforcement Learning" }, "github": { "value": 27, "source": "GitHub", "sourceUrl": "https://github.com/TTomilin/COOM/stargazers", "repository": "TTomilin/COOM", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-b02", "name": "HASARD", "title": "HASARD: A Benchmark for Vision-Based Safe Reinforcement Learning in Embodied Agents", "domains": [ "game" ], "kinds": [ "papers", "benchmarks" ], "year": 2025, "venue": "ICLR 2025", "summary": "Separates task rewards from violation costs to evaluate visual agents under hazards such as dangerous terrain and friendly fire.", "tags": [ "Safety & verification", "Benchmarks" ], "environment": "Doom / ViZDoom.", "interface": "Partial visual observations feed constrained PPO variants with separate rewards and costs, using full or simplified action configurations.", "limits": "An expected-cost budget does not guarantee zero violations in every trial. Fixed penalties differ from constrained optimization; action spaces, budgets, and curriculum resources must be aligned. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5ab848771ff8c9c47aac4128e2ef9f4e-Abstract-Conference.html", "project": "https://sites.google.com/view/hasard-bench/", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5ab848771ff8c9c47aac4128e2ef9f4e-Paper-Conference.pdf" }, "rank": 14, "sourceId": "B02", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/5ab848771ff8c9c47aac4128e2ef9f4e-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Tristan Tomilin", "Meng Fang", "Mykola Pechenizkiy" ], "date": "2025-05-01", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-b02,\n title = {HASARD: A Benchmark for Vision-Based Safe Reinforcement Learning in Embodied Agents},\n author = {Tristan Tomilin and Meng Fang and Mykola Pechenizkiy},\n year = {2025},\n url = {https://proceedings.iclr.cc/paper_files/paper/2025/hash/5ab848771ff8c9c47aac4128e2ef9f4e-Abstract-Conference.html}\n}", "preview": "assets/papers/kb-b02.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2025/file/5ab848771ff8c9c47aac4128e2ef9f4e-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4414577619", "updatedAt": "2026-10-19", "recordTitle": "HASARD: A Benchmark for Vision-Based Safe Reinforcement Learning in Embodied Agents" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-w01", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "GameNGen", "title": "Diffusion Models Are Real-Time Game Engines", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "ICLR 2025", "summary": "Generates interactive DOOM frames conditioned on past images and player actions, exploring neural networks as game simulators.", "tags": [ "World models", "Environment generation" ], "environment": "DOOM / ViZDoom.", "interface": "Past game frames and player actions condition diffusion-based next-frame prediction. PPO primarily collects training trajectories.", "limits": "Generation frame rate, visual similarity, and control success are distinct. Single-frame teacher-forcing metrics do not establish long-horizon logical consistency. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b71ecea210f7159f31e46631fe5c838f-Abstract-Conference.html", "project": "https://gamengen.github.io/", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b71ecea210f7159f31e46631fe5c838f-Paper-Conference.pdf" }, "rank": 9, "sourceId": "W01", "sourceEvidence": "Spot checks of key claims", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2025/hash/b71ecea210f7159f31e46631fe5c838f-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Dani Valevski", "Yaniv Leviathan", "Moab Arar", "Shlomi Fruchter" ], "date": "2025-05-01", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-w01,\n title = {Diffusion Models Are Real-Time Game Engines},\n author = {Dani Valevski and Yaniv Leviathan and Moab Arar and Shlomi Fruchter},\n year = {2025},\n url = {https://proceedings.iclr.cc/paper_files/paper/2025/hash/b71ecea210f7159f31e46631fe5c838f-Abstract-Conference.html}\n}", "preview": "assets/papers/kb-w01.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2025/file/b71ecea210f7159f31e46631fe5c838f-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 7, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402704721", "updatedAt": "2026-10-19", "recordTitle": "Diffusion Models Are Real-Time Game Engines" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-w02", "name": "DIAMOND", "title": "Diffusion for World Modeling: Visual Details Matter in Atari", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "NeurIPS 2024", "summary": "Preserves visual detail in diffusion world models for policy learning in Atari, with a separate interactive CS:GO frame-generation extension.", "tags": [ "World models", "Diffusion models" ], "environment": "Atari; Counter-Strike: Global Offensive, Dust II.", "interface": "Human CS:GO frames and actions feed low-resolution dynamics and super-resolution generation. The shooter extension does not train an RL policy.", "limits": "Separate Atari policy learning from CS:GO generation demos. The shooter section was added after acceptance; Atari benchmark scores are not shooter performance. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/6bdde0373d53d4a501249547084bed43-Abstract-Conference.html", "code": "https://github.com/eloialonso/diamond", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2024/file/6bdde0373d53d4a501249547084bed43-Paper-Conference.pdf" }, "rank": 29, "sourceId": "W02", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS", "Multi-game" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/6bdde0373d53d4a501249547084bed43-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Eloi Alonso", "Adam Jelley", "Vincent Micheli", "Anssi Kanervisto", "Amos Storkey", "Tim Pearce", "François Fleuret" ], "date": "2024-12-16", "dateLabel": "Publisher record date", "doi": "10.52202/079017-1873", "bibtex": "@misc{kb-w02,\n title = {Diffusion for World Modeling: Visual Details Matter in Atari},\n author = {Eloi Alonso and Adam Jelley and Vincent Micheli and Anssi Kanervisto and Amos Storkey and Tim Pearce and François Fleuret},\n year = {2024},\n url = {https://proceedings.neurips.cc/paper_files/paper/2024/hash/6bdde0373d53d4a501249547084bed43-Abstract-Conference.html},\n doi = {10.52202/079017-1873}\n}", "preview": "assets/papers/kb-w02.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2024/file/6bdde0373d53d4a501249547084bed43-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4398229893", "updatedAt": "2026-10-12", "recordTitle": "Diffusion for World Modeling: Visual Details Matter in Atari" }, "github": { "value": 2122, "source": "GitHub", "sourceUrl": "https://github.com/eloialonso/diamond/stargazers", "repository": "eloialonso/diamond", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-w03", "name": "StateSpaceDiffuser", "title": "StateSpaceDiffuser: Bringing Long Context to Diffusion World Models", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "NeurIPS 2025", "summary": "Compresses long histories with a state-space sequence model to give diffusion world models memory when revisiting scenes.", "tags": [ "World models", "Planning & memory" ], "environment": "MiniGrid; Counter-Strike: Global Offensive.", "interface": "Past images and actions pass through Cosmos encoding and Mamba memory to condition DIAMOND generation, including constructed reverse trajectories.", "limits": "Forward/reverse trajectories diagnose memory rather than physically reversible dynamics. MiniGrid metrics and CS:GO user preferences differ; neither is a competitive win rate. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.nips.cc/paper_files/paper/2025/hash/63943ee9fe347f3d95892cf87d9a42e6-Abstract-Conference.html", "code": "https://github.com/insait-institute/StateSpaceDiffuser", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2025/file/63943ee9fe347f3d95892cf87d9a42e6-Paper-Conference.pdf" }, "rank": 30, "sourceId": "W03", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.nips.cc/paper_files/paper/2025/hash/63943ee9fe347f3d95892cf87d9a42e6-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Nedko Savov", "Naser Kazemi", "Deheng Zhang", "Danda Pani Paudel", "Xi Wang", "Luc V Gool" ], "date": "2026-04-23", "dateLabel": "Publisher record date", "doi": "10.52202/085713-2316", "bibtex": "@misc{kb-w03,\n title = {StateSpaceDiffuser: Bringing Long Context to Diffusion World Models},\n author = {Nedko Savov and Naser Kazemi and Deheng Zhang and Danda Pani Paudel and Xi Wang and Luc V Gool},\n year = {2026},\n url = {https://proceedings.nips.cc/paper_files/paper/2025/hash/63943ee9fe347f3d95892cf87d9a42e6-Abstract-Conference.html},\n doi = {10.52202/085713-2316}\n}", "preview": "assets/papers/kb-w03.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2025/file/63943ee9fe347f3d95892cf87d9a42e6-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4416048054", "updatedAt": "2026-10-19", "recordTitle": "StateSpaceDiffuser: Bringing Long Context to Diffusion World Models" }, "github": { "value": 6, "source": "GitHub", "sourceUrl": "https://github.com/insait-institute/StateSpaceDiffuser/stargazers", "repository": "insait-institute/StateSpaceDiffuser", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-s01", "name": "EDELINE", "title": "EDELINE: Enhancing Memory in Diffusion-based World Models via Linear-Time Sequence Modeling", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "NeurIPS 2025", "summary": "Adds linear-time sequence memory to a diffusion world model, predicting images, rewards, and termination for policy learning in imagined trajectories.", "tags": [ "World models", "Planning & memory" ], "environment": "ViZDoom Basic, DeadlyCorridor, DefendCenter, HealthGathering, and PredictPosition; also Atari, Crafter, and MiniGrid.", "interface": "Recent images and actions plus recurrent Mamba memory predict diffusion frames and shared reward/termination outputs for imagined policy learning.", "limits": "Distinguish image MSE from real policy returns. The five Doom tasks use different metrics; linear sequence cost does not imply real-time generation for the complete system. Compiled from existing paper notes; not independently reproduced. Original review claims are retained as source information.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/97c903fbf21a7d863af2015d8803ca8f-Abstract-Conference.html", "code": "https://github.com/LJH-coding/EDELINE", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2025/file/97c903fbf21a7d863af2015d8803ca8f-Paper-Conference.pdf" }, "rank": 31, "sourceId": "S01", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/97c903fbf21a7d863af2015d8803ca8f-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Jia-Hua Lee", "Bor-Jiun Lin", "Wei-Fang Sun", "Chun-Yi Lee" ], "date": "2026-04-23", "dateLabel": "Publisher record date", "doi": "10.52202/085713-3526", "bibtex": "@misc{kb-s01,\n title = {EDELINE: Enhancing Memory in Diffusion-based World Models via Linear-Time Sequence Modeling},\n author = {Jia-Hua Lee and Bor-Jiun Lin and Wei-Fang Sun and Chun-Yi Lee},\n year = {2026},\n url = {https://proceedings.neurips.cc/paper_files/paper/2025/hash/97c903fbf21a7d863af2015d8803ca8f-Abstract-Conference.html},\n doi = {10.52202/085713-3526}\n}", "preview": "assets/papers/kb-s01.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2025/file/97c903fbf21a7d863af2015d8803ca8f-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4407123592", "updatedAt": "2026-10-12", "recordTitle": "EDELINE: Enhancing Memory in Diffusion-based World Models via Linear-Time Sequence Modeling" }, "github": { "value": 20, "source": "GitHub", "sourceUrl": "https://github.com/LJH-coding/EDELINE/stargazers", "repository": "LJH-coding/EDELINE", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-t01", "name": "Contextual Similarity Distillation", "title": "Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "ICLR 2026 Main Conference", "summary": "Approximates ensemble uncertainty with a single model and uses it as an exploration signal for sparse visual navigation.", "tags": [ "Exploration & skills", "Representation learning" ], "environment": "Three starting-position and sparse-reward variants of ViZDoom MyWayHome maze navigation.", "interface": "First-person maze images feed goal-reaching navigation actions. The exact action set needs verification; there is no shooting or opponent combat.", "limits": "Stable goal discovery does not mean every episode succeeds. Distinguish author manuscripts from the official version; navigation improvements do not establish full FPS control.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://iclr.cc/virtual/2026/poster/10008663", "code": "https://github.com/anyboby/contextual-similarity-distillation", "pdf": "https://arxiv.org/pdf/2503.11339" }, "rank": 32, "sourceId": "T01", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4417285496", "checkedAt": "2026-09-14", "authors": [ "Moritz A. Zanger", "Pascal R. Van der Vaart", "Wendelin Böhmer", "Matthijs T. J. Spaan" ], "date": "2025-03-14", "dateLabel": "Indexed publication date (OpenAlex)", "arxivId": "2503.11339", "bibtex": "@misc{kb-t01,\n title = {Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model},\n author = {Moritz A. Zanger and Pascal R. Van der Vaart and Wendelin Böhmer and Matthijs T. J. Spaan},\n year = {2025},\n url = {https://iclr.cc/virtual/2026/poster/10008663}\n}", "preview": "assets/papers/kb-t01.png", "previewSourceUrl": "https://arxiv.org/pdf/2503.11339" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4417285496", "updatedAt": "2026-10-19", "recordTitle": "Contextual Similarity Distillation: Ensemble Uncertainties with a Single Model" }, "github": { "value": 1, "source": "GitHub", "sourceUrl": "https://github.com/anyboby/contextual-similarity-distillation/stargazers", "repository": "anyboby/contextual-similarity-distillation", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-t02", "name": "RATE", "title": "Recurrent Action Transformer with Memory", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "ICLR 2026 Main Conference", "summary": "Adds memory across segments to a decision transformer for retaining clues and predicting return-conditioned actions in long-horizon tasks.", "tags": [ "Planning & memory", "Imitation learning" ], "environment": "ViZDoom-Two-Colors; also T-Maze, Memory Maze, MiniGrid, POPGym, Atari, and MuJoCo.", "interface": "First-person images, history, and return conditioning support remembering a color and navigating to collect an item. The exact action set needs verification; no aiming or shooting is involved.", "limits": "Separate benchmarks and cue-disappearance settings. The longest memory span in another environment is not a Doom result; record offline-data conditions and training/test lengths together.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://iclr.cc/virtual/2026/poster/10007816", "project": "https://sites.google.com/view/rate-model/", "pdf": "https://arxiv.org/pdf/2306.09459" }, "rank": 33, "sourceId": "T02", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4381245688", "checkedAt": "2026-09-14", "authors": [ "Egor Cherepanov", "Alexey Staroverov", "Alexey K. Kovalev", "Aleksandr I. Panov" ], "date": "2023-06-15", "dateLabel": "Indexed publication date (OpenAlex)", "arxivId": "2306.09459", "bibtex": "@misc{kb-t02,\n title = {Recurrent Action Transformer with Memory},\n author = {Egor Cherepanov and Alexey Staroverov and Alexey K. Kovalev and Aleksandr I. Panov},\n year = {2023},\n url = {https://iclr.cc/virtual/2026/poster/10007816}\n}", "preview": "assets/papers/kb-t02.png", "previewSourceUrl": "https://arxiv.org/pdf/2306.09459" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4381245688", "updatedAt": "2026-10-19", "recordTitle": "Recurrent Action Transformer with Memory" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c01", "name": "VisionMask", "title": "Why the Agent Made that Decision: Contrastive Explanation Learning for Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "IJCAI 2025", "summary": "Learns contrastive visual explanations between actions to inspect why a reinforcement-learning policy makes a decision.", "tags": [ "Safety & verification", "Representation learning" ], "environment": "ViZDoom; explanations of first-person visual DQN decisions.", "interface": "ViZDoom images and a frozen DQN produce visual masks for candidate actions. The control policy is not updated.", "limits": "Accuracy measures action agreement after masking, not win rate. Explanation metrics have tradeoffs; access to policy gradients means this is not a purely black-box method.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://www.ijcai.org/proceedings/2025/74", "pdf": "https://www.ijcai.org/proceedings/2025/0074.pdf" }, "rank": 34, "sourceId": "C01", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.ijcai.org/proceedings/2025/74", "checkedAt": "2026-09-14", "authors": [ "Rui Zuo", "Simon Khan", "Zifan Wang", "Garrett Ethan Katz", "Qinru Qiu" ], "date": "2025-09-16", "dateLabel": "Publisher record date", "doi": "10.24963/ijcai.2025/74", "bibtex": "@misc{kb-c01,\n title = {Why the Agent Made that Decision: Contrastive Explanation Learning for Reinforcement Learning},\n author = {Rui Zuo and Simon Khan and Zifan Wang and Garrett Ethan Katz and Qinru Qiu},\n year = {2025},\n url = {https://www.ijcai.org/proceedings/2025/74},\n doi = {10.24963/ijcai.2025/74}\n}", "preview": "assets/papers/kb-c01.png", "previewSourceUrl": "https://www.ijcai.org/proceedings/2025/0074.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4414360152", "updatedAt": "2026-10-19", "recordTitle": "Why the Agent Made that Decision: Contrastive Explanation Learning for Reinforcement Learning" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c02", "name": "EVAPS", "title": "Enhancing Robot Program Synthesis Through Environmental Context", "domains": [ "game", "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Aligns program structure with visual context before and after execution to repair candidate programs in a robot domain-specific language.", "tags": [ "Programs & tools", "Planning & memory" ], "environment": "ViZDoom short-program synthesis with movement, turning, and attack actions.", "interface": "Input/output examples, candidate program graphs, and before/after images produce short DSL programs with movement, turning, attack, and enemy-perception conditions.", "limits": "Program generalization accuracy is not combat win rate. Separate top-1 results from multi-candidate budgets; a restricted DSL, perception predicates, and short executions do not establish full autonomous FPS play.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/0c1e94af650f5c74b1f3da467c2308c2-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/0c1e94af650f5c74b1f3da467c2308c2-Paper-Conference.pdf" }, "rank": 35, "sourceId": "C02", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/0c1e94af650f5c74b1f3da467c2308c2-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Tianyi Chen", "Qidi Wang", "Zhen Dong", "Liwei Shen", "Xin Peng" ], "date": "2023-12-15", "dateLabel": "Publisher record date", "doi": "10.52202/075280-0171", "bibtex": "@misc{kb-c02,\n title = {Enhancing Robot Program Synthesis Through Environmental Context},\n author = {Tianyi Chen and Qidi Wang and Zhen Dong and Liwei Shen and Xin Peng},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/0c1e94af650f5c74b1f3da467c2308c2-Abstract-Conference.html},\n doi = {10.52202/075280-0171}\n}", "preview": "assets/papers/kb-c02.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/0c1e94af650f5c74b1f3da467c2308c2-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7133226547", "updatedAt": "2026-10-19", "recordTitle": "Enhancing Robot Program Synthesis Through Environmental Context" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c03", "name": "Just Cluster It", "title": "Just Cluster It: An Approach for Exploration in High-Dimensions using Clustering and Pre-Trained Representations", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICML 2024", "summary": "Approximates state visitation counts by clustering visual representations, providing exploration bonuses within and across episodes.", "tags": [ "Exploration & skills", "Representation learning" ], "environment": "ViZDoom MyWayHome Sparse and Very Sparse; non-combat navigation.", "interface": "First-person images use fixed random or pretrained encoders; PPO performs object-seeking navigation. The exact action set needs verification.", "limits": "Separate random and pretrained representations. DINO is not uniformly best on Doom; baselines use different underlying algorithms and relatively few seeds. Navigation convergence is not shooting ability.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v235/wagner24a.html", "code": "https://github.com/stefanwm13/just-cluster-it-public", "pdf": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/wagner24a/wagner24a.pdf" }, "rank": 36, "sourceId": "C03", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v235/wagner24a.html", "checkedAt": "2026-09-14", "authors": [ "Stefan Sylvius Wagner", "Stefan Harmeling" ], "date": "2024-07-08", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-c03,\n title = {Just Cluster It: An Approach for Exploration in High-Dimensions using Clustering and Pre-Trained Representations},\n author = {Stefan Sylvius Wagner and Stefan Harmeling},\n year = {2024},\n url = {https://proceedings.mlr.press/v235/wagner24a.html}\n}", "preview": "assets/papers/kb-c03.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v235/main/assets/wagner24a/wagner24a.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4391622581", "updatedAt": "2026-10-19", "recordTitle": "Just Cluster It: An Approach for Exploration in High-Dimensions using Clustering and Pre-Trained Representations" }, "github": { "value": 4, "source": "GitHub", "sourceUrl": "https://github.com/stefanwm13/just-cluster-it-public/stargazers", "repository": "stefanwm13/just-cluster-it-public", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c04", "name": "E3B", "title": "Exploration via Elliptical Episodic Bonuses", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "NeurIPS 2022", "summary": "Constructs elliptical episodic exploration bonuses from feature coverage, using inverse dynamics to learn controllable state representations.", "tags": [ "Exploration & skills", "Reinforcement learning" ], "environment": "ViZDoom MyWayHome Dense, Sparse, and Very Sparse; non-combat navigation.", "interface": "Visual states encoded through inverse dynamics generate intrinsic rewards for navigation. The exact action set needs verification; there is no shooting.", "limits": "Broad advantages on MiniHack do not imply uniform superiority on Doom. Fixed maps differ from changing environments; compare episodic bonuses, representations, and compute costs together.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/f4f79698d48bdc1a6dec20583724182b-Abstract-Conference.html", "project": "https://e3bagent.github.io/", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2022/file/f4f79698d48bdc1a6dec20583724182b-Paper-Conference.pdf" }, "rank": 37, "sourceId": "C04", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/f4f79698d48bdc1a6dec20583724182b-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Mikael Henaff", "Roberta Raileanu", "Minqi Jiang", "Tim Rocktäschel" ], "date": "2022-12-06", "dateLabel": "Publisher record date", "doi": "10.52202/068431-2728", "bibtex": "@misc{kb-c04,\n title = {Exploration via Elliptical Episodic Bonuses},\n author = {Mikael Henaff and Roberta Raileanu and Minqi Jiang and Tim Rocktäschel},\n year = {2022},\n url = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/f4f79698d48bdc1a6dec20583724182b-Abstract-Conference.html},\n doi = {10.52202/068431-2728}\n}", "preview": "assets/papers/kb-c04.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/file/f4f79698d48bdc1a6dec20583724182b-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 8, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4306176261", "updatedAt": "2026-10-12", "recordTitle": "Exploration via Elliptical Episodic Bonuses" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c05", "name": "Successor Feature Landmarks", "title": "Successor Feature Landmarks for Long-Horizon Goal-Conditioned Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2021, "venue": "NeurIPS 2021", "summary": "Organizes reachable landmarks with successor features, combining graph planning and low-level goal policies for long-horizon navigation.", "tags": [ "Planning & memory", "Exploration & skills" ], "environment": "Large ViZDoom visual mazes: SGM Map and Test 1/2/4/6; non-combat navigation.", "interface": "First-person images use pretrained SPTM representations; a high-level planner chooses landmarks and a low-level policy executes discrete navigation actions.", "limits": "Separate fixed/random spawn points, difficulty levels, and maps. Training steps omit some pretraining costs; strong results on difficult tasks do not imply superiority on every task.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2021/hash/e27c71957d1e6c223e0d48a165da2ee1-Abstract.html", "code": "https://github.com/2016choang/sfl", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2021/file/e27c71957d1e6c223e0d48a165da2ee1-Paper.pdf" }, "rank": 38, "sourceId": "C05", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "publicationDate": "2021-12", "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2021/hash/e27c71957d1e6c223e0d48a165da2ee1-Abstract.html", "checkedAt": "2026-09-14", "authors": [ "Christopher Hoang", "Sungryull Sohn", "Jongwook Choi", "Wilka Carvalho", "Honglak Lee" ], "date": "2021-12-06", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-c05,\n title = {Successor Feature Landmarks for Long-Horizon Goal-Conditioned Reinforcement Learning},\n author = {Christopher Hoang and Sungryull Sohn and Jongwook Choi and Wilka Carvalho and Honglak Lee},\n year = {2021},\n url = {https://proceedings.neurips.cc/paper_files/paper/2021/hash/e27c71957d1e6c223e0d48a165da2ee1-Abstract.html}\n}", "preview": "assets/papers/kb-c05.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2021/file/e27c71957d1e6c223e0d48a165da2ee1-Paper.pdf" }, "metrics": { "citations": { "value": 11, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3212475937", "updatedAt": "2026-10-19", "recordTitle": "Successor Feature Landmarks for Long-Horizon Goal-Conditioned Reinforcement Learning" }, "github": { "value": 8, "source": "GitHub", "sourceUrl": "https://github.com/2016choang/sfl/stargazers", "repository": "2016choang/sfl", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c06", "name": "CoSHRL", "title": "Handling Long and Richly Constrained Tasks through Constrained Hierarchical Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "AAAI 2024 — Safe, Robust and Responsible AI Track", "summary": "Combines goal-conditioned skills and constrained path search to balance goal reaching against accumulated hazard costs over long tasks.", "tags": [ "Planning & memory", "Safety & verification" ], "environment": "Safe ViZDoom four-leaf-clover maze with hazardous floors; non-combat navigation.", "interface": "The exact state and visual processing need verification. Low-level actions move in four absolute directions; the high-level policy chooses constrained subgoals.", "limits": "Goal-reaching and constraint-satisfaction rates are distinct. Stochastic CVaR evidence comes from environments other than Doom. Nonzero violation rates preclude absolute safety claims; the action interface is simplified.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://ojs.aaai.org/index.php/AAAI/article/view/30132", "pdf": "https://ojs.aaai.org/index.php/AAAI/article/download/30132/32004" }, "rank": 39, "sourceId": "C06", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4393157199", "checkedAt": "2026-09-14", "authors": [ "Yuxiao Lu", "Arunesh Sinha", "Pradeep Varakantham" ], "date": "2024-03-24", "dateLabel": "Indexed publication date (OpenAlex)", "bibtex": "@misc{kb-c06,\n title = {Handling Long and Richly Constrained Tasks through Constrained Hierarchical Reinforcement Learning},\n author = {Yuxiao Lu and Arunesh Sinha and Pradeep Varakantham},\n year = {2024},\n url = {https://ojs.aaai.org/index.php/AAAI/article/view/30132}\n}", "preview": "assets/papers/kb-c06.png", "previewSourceUrl": "https://ojs.aaai.org/index.php/AAAI/article/download/30132/32004" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4393157199", "updatedAt": "2026-10-19", "recordTitle": "Handling Long and Richly Constrained Tasks through Constrained Hierarchical Reinforcement Learning" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c08", "name": "PE-DQN", "title": "Diverse Projection Ensembles for Distributional Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICLR 2024", "summary": "Ensembles different return-distribution representations and uses disagreement to estimate epistemic uncertainty for exploration.", "tags": [ "Exploration & skills", "Reinforcement learning" ], "environment": "ViZDoom MyWayHome Dense, Sparse, and Very Sparse; non-combat navigation.", "interface": "Downsampled grayscale images and stacked frames map to navigation actions. The exact action set needs verification; no shooting or opponents are present.", "limits": "All training seeds discovering the goal does not imply 100% success in every episode. Distributional return modeling is not distributed computing; align reward sparsity and exploration parameters.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/cfea00ddf5a5f514bd3368d639d1933d-Abstract-Conference.html", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2024/file/cfea00ddf5a5f514bd3368d639d1933d-Paper-Conference.pdf" }, "rank": 40, "sourceId": "C08", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/cfea00ddf5a5f514bd3368d639d1933d-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Moritz Akiya Zanger", "Wendelin Boehmer", "Matthijs T. J. Spaan" ], "date": "2024-05-31", "dateLabel": "Publisher record date", "bibtex": "@misc{kb-c08,\n title = {Diverse Projection Ensembles for Distributional Reinforcement Learning},\n author = {Moritz Akiya Zanger and Wendelin Boehmer and Matthijs T. J. Spaan},\n year = {2024},\n url = {https://proceedings.iclr.cc/paper_files/paper/2024/hash/cfea00ddf5a5f514bd3368d639d1933d-Abstract-Conference.html}\n}", "preview": "assets/papers/kb-c08.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/file/cfea00ddf5a5f514bd3368d639d1933d-Paper-Conference.pdf" }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "kb-c09", "name": "MOSS", "title": "A Mixture of Surprises for Unsupervised Reinforcement Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "NeurIPS 2022", "summary": "Mixes novelty-seeking and unpredictability-reducing skills during unsupervised pretraining, then adapts to downstream control tasks.", "tags": [ "Exploration & skills", "Reinforcement learning" ], "environment": "ViZDoom Defend The Line shooting and Take Cover projectile-avoidance survival.", "interface": "The precise observation encoding needs verification. Skill-conditioned policies handle shooting and avoidance, followed by task-reward fine-tuning.", "limits": "Reported task scores are not kills or win rates. The paper uses interquartile means and standard errors rather than ordinary means and standard deviations. Unsupervised pretraining does not imply zero-shot deployment.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/a7667ee5d545a43d2f0fda98863c260e-Abstract-Conference.html", "code": "https://github.com/LeapLabTHU/MOSS", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2022/file/a7667ee5d545a43d2f0fda98863c260e-Paper-Conference.pdf" }, "rank": 41, "sourceId": "C09", "sourceEvidence": "Compiled from original notes", "selectionReason": "Published within the five-year window; official paper page checked.", "gameTypes": [ "FPS" ], "selectionTrack": "peer-reviewed", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/a7667ee5d545a43d2f0fda98863c260e-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Andrew Zhao", "Matthieu Lin", "Yangguang Li", "Yong-jin Liu", "Gao Huang" ], "date": "2022-12-06", "dateLabel": "Publisher record date", "doi": "10.52202/068431-1891", "bibtex": "@misc{kb-c09,\n title = {A Mixture of Surprises for Unsupervised Reinforcement Learning},\n author = {Andrew Zhao and Matthieu Lin and Yangguang Li and Yong-jin Liu and Gao Huang},\n year = {2022},\n url = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/a7667ee5d545a43d2f0fda98863c260e-Abstract-Conference.html},\n doi = {10.52202/068431-1891}\n}", "preview": "assets/papers/kb-c09.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/file/a7667ee5d545a43d2f0fda98863c260e-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4306295186", "updatedAt": "2026-10-12", "recordTitle": "A Mixture of Surprises for Unsupervised Reinforcement Learning" }, "github": { "value": 23, "source": "GitHub", "sourceUrl": "https://github.com/LeapLabTHU/MOSS/stargazers", "repository": "LeapLabTHU/MOSS", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "mp5", "name": "MP5", "title": "MP5: A Multi-modal Open-ended Embodied System in Minecraft via Active Perception", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "CVPR 2024", "summary": "Combines active perception and modular planning to decompose and execute Minecraft tasks with contextual and procedural dependencies.", "tags": [ "Planning & memory", "Language & action" ], "environment": "Minecraft.", "interface": "Multimodal observations, active perception, and modular action execution.", "limits": "Evaluated with specified Minecraft interfaces and tasks; results do not directly establish transfer to real robots.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://openaccess.thecvf.com/content/CVPR2024/html/Qin_MP5_A_Multi-modal_Open-ended_Embodied_System_in_Minecraft_via_Active_CVPR_2024_paper.html", "project": "https://iranqin.github.io/MP5.github.io/", "code": "https://github.com/IranQin/MP5", "pdf": "https://openaccess.thecvf.com/content/CVPR2024/papers/Qin_MP5_A_Multi-modal_Open-ended_Embodied_System_in_Minecraft_via_Active_CVPR_2024_paper.pdf" }, "rank": 11, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://openaccess.thecvf.com/content/CVPR2024/html/Qin_MP5_A_Multi-modal_Open-ended_Embodied_System_in_Minecraft_via_Active_CVPR_2024_paper.html", "checkedAt": "2026-09-14", "authors": [ "Yiran Qin", "Enshen Zhou", "Qichang Liu", "Zhenfei Yin", "Lu Sheng", "Ruimao Zhang", "Yu Qiao", "Jing Shao" ], "date": "2024", "dateLabel": "Publisher record date", "bibtex": "@misc{mp5,\n title = {MP5: A Multi-modal Open-ended Embodied System in Minecraft via Active Perception},\n author = {Yiran Qin and Enshen Zhou and Qichang Liu and Zhenfei Yin and Lu Sheng and Ruimao Zhang and Yu Qiao and Jing Shao},\n year = {2024},\n url = {https://openaccess.thecvf.com/content/CVPR2024/html/Qin_MP5_A_Multi-modal_Open-ended_Embodied_System_in_Minecraft_via_Active_CVPR_2024_paper.html}\n}", "preview": "assets/papers/mp5.png", "previewSourceUrl": "https://openaccess.thecvf.com/content/CVPR2024/papers/Qin_MP5_A_Multi-modal_Open-ended_Embodied_System_in_Minecraft_via_Active_CVPR_2024_paper.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4389761566", "updatedAt": "2026-10-12", "recordTitle": "MP5: A Multi-modal Open-ended Embodied System in Minecraft via Active Perception" }, "github": { "value": 108, "source": "GitHub", "sourceUrl": "https://github.com/IranQin/MP5/stargazers", "repository": "IranQin/MP5", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "deps", "name": "DEPS", "title": "Describe, Explain, Plan and Select: Interactive Planning with LLMs Enables Open-World Multi-Task Agents", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Connects execution descriptions, failure explanations, and subgoal selection to language planning, revising long-horizon Minecraft plans with feedback.", "tags": [ "Planning & memory", "Language & action" ], "environment": "Minecraft.", "interface": "Language planning and skill-level execution.", "limits": "Depends on the task interface and low-level execution capabilities. Evaluate plan feasibility separately from complete real-time control.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/6b8dfb8c0c12e6fafc6c256cb08a5ca7-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6b8dfb8c0c12e6fafc6c256cb08a5ca7-Paper-Conference.pdf" }, "rank": 43, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/6b8dfb8c0c12e6fafc6c256cb08a5ca7-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Zihao Wang", "Shaofei Cai", "Guanzhou Chen", "Anji Liu", "Xiaojian (Shawn) Ma", "Yitao Liang" ], "date": "2023-12-15", "dateLabel": "Publisher record date", "doi": "10.52202/075280-1480", "bibtex": "@misc{deps,\n title = {Describe, Explain, Plan and Select: Interactive Planning with LLMs Enables Open-World Multi-Task Agents},\n author = {Zihao Wang and Shaofei Cai and Guanzhou Chen and Anji Liu and Xiaojian (Shawn) Ma and Yitao Liang},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/6b8dfb8c0c12e6fafc6c256cb08a5ca7-Abstract-Conference.html},\n doi = {10.52202/075280-1480}\n}", "preview": "assets/papers/deps.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/6b8dfb8c0c12e6fafc6c256cb08a5ca7-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 55, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4319323461", "updatedAt": "2026-10-19", "recordTitle": "Describe, Explain, Plan and Select: Interactive Planning with Large Language Models Enables Open-World Multi-Task Agents" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "voyager", "name": "Voyager", "title": "Voyager: An Open-Ended Embodied Agent with Large Language Models", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "TMLR 2024", "summary": "Combines an automatic curriculum, reusable code skills, and execution feedback for continuing exploration and skill acquisition in Minecraft.", "tags": [ "Programs & tools", "Exploration & skills" ], "environment": "Minecraft.", "interface": "JavaScript skill code, game state, and execution feedback.", "limits": "Depends on program interfaces such as Mineflayer and language-model services; it does not operate solely through pixels, keyboard, and mouse.", "status": "selected", "evidence": "Official paper PDF and author publication list checked; OpenReview required browser verification. Experiments not reproduced.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=P8E4Br72j3", "code": "https://github.com/MineDojo/Voyager", "project": "https://voyager.minedojo.org/", "pdf": "https://arxiv.org/pdf/2305.16291" }, "rank": 3, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2305.16291", "checkedAt": "2026-09-14", "authors": [ "Guanzhi Wang", "Yuqi Xie", "Yunfan Jiang", "Ajay Mandlekar", "Chaowei Xiao", "Yuke Zhu", "Linxi Fan", "Anima Anandkumar" ], "date": "2023-05-25", "dateLabel": "First arXiv submission", "arxivId": "2305.16291", "bibtex": "@misc{voyager,\n title = {Voyager: An Open-Ended Embodied Agent with Large Language Models},\n author = {Guanzhi Wang and Yuqi Xie and Yunfan Jiang and Ajay Mandlekar and Chaowei Xiao and Yuke Zhu and Linxi Fan and Anima Anandkumar},\n year = {2023},\n url = {https://openreview.net/forum?id=P8E4Br72j3}\n}", "preview": "assets/papers/voyager.png", "previewSourceUrl": "https://arxiv.org/pdf/2305.16291" }, "metrics": { "citations": { "value": 221, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4378505261", "updatedAt": "2026-10-19", "recordTitle": "Voyager: An Open-Ended Embodied Agent with Large Language Models" }, "github": { "value": 7271, "source": "GitHub", "sourceUrl": "https://github.com/MineDojo/Voyager/stargazers", "repository": "MineDojo/Voyager", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "vpt", "institutions": [ { "name": "OpenAI", "url": "https://openai.com/", "logo": "assets/institutions/openai.png", "mark": "OAI" } ], "name": "Video PreTraining", "title": "Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "NeurIPS 2022", "summary": "Learns inverse dynamics from a small action-labeled dataset, then annotates internet videos to scale Minecraft behavioral pretraining.", "tags": [ "Imitation learning", "Representation learning" ], "environment": "Minecraft.", "interface": "Game images map to keyboard and mouse actions.", "limits": "Record internet-video pretraining separately from downstream task fine-tuning. The policy is not universally deployable without environment adaptation.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/9c7008aff45b5d8f0973b23e1a22ada0-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2022/file/9c7008aff45b5d8f0973b23e1a22ada0-Paper-Conference.pdf" }, "rank": 15, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/9c7008aff45b5d8f0973b23e1a22ada0-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Bowen Baker", "Ilge Akkaya", "Peter Zhokov", "Joost Huizinga", "Jie Tang", "Adrien Ecoffet", "Brandon Houghton", "Raul Sampedro", "Jeff Clune" ], "date": "2022-12-06", "dateLabel": "Publisher record date", "doi": "10.52202/068431-1789", "bibtex": "@misc{vpt,\n title = {Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos},\n author = {Bowen Baker and Ilge Akkaya and Peter Zhokov and Joost Huizinga and Jie Tang and Adrien Ecoffet and Brandon Houghton and Raul Sampedro and Jeff Clune},\n year = {2022},\n url = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/9c7008aff45b5d8f0973b23e1a22ada0-Abstract-Conference.html},\n doi = {10.52202/068431-1789}\n}", "preview": "assets/papers/vpt.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/file/9c7008aff45b5d8f0973b23e1a22ada0-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 48, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4283460722", "updatedAt": "2026-10-12", "recordTitle": "Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "steve1", "name": "STEVE-1", "title": "STEVE-1: A Generative Model for Text-to-Behavior in Minecraft", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Conditions a VPT behavior model on text and visual goals to follow short-horizon, open-ended instructions in Minecraft.", "tags": [ "Language & action", "Imitation learning" ], "environment": "Minecraft.", "interface": "Text or visual goals condition low-level behavior.", "limits": "Short-horizon instruction following is the main scope. Complex long-horizon tasks still require planning and memory.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/dd03f856fc7f2efeec8b1c796284561d-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/dd03f856fc7f2efeec8b1c796284561d-Paper-Conference.pdf" }, "rank": 46, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/dd03f856fc7f2efeec8b1c796284561d-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Shalev Lifshitz", "Keiran Paster", "Harris Chan", "Jimmy Ba", "Sheila McIlraith" ], "date": "2023-12-15", "dateLabel": "Publisher record date", "doi": "10.52202/075280-3064", "bibtex": "@misc{steve1,\n title = {STEVE-1: A Generative Model for Text-to-Behavior in Minecraft},\n author = {Shalev Lifshitz and Keiran Paster and Harris Chan and Jimmy Ba and Sheila McIlraith},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/dd03f856fc7f2efeec8b1c796284561d-Abstract-Conference.html},\n doi = {10.52202/075280-3064}\n}", "preview": "assets/papers/steve1.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/dd03f856fc7f2efeec8b1c796284561d-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 7, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4379357893", "updatedAt": "2026-09-28", "recordTitle": "STEVE-1: A Generative Model for Text-to-Behavior in Minecraft" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "openvla", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" }, { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" }, { "name": "Toyota Research Institute", "url": "https://www.tri.global/", "logo": "assets/institutions/toyota-research-institute.svg", "mark": "TRI" } ], "name": "OpenVLA", "title": "OpenVLA: An Open-Source Vision-Language-Action Model", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "CoRL 2024", "summary": "Unifies vision, language, and robot actions in an open model, with released weights and training tools for adaptation to new tasks.", "tags": [ "Vision-language-action", "Robot learning" ], "environment": "Multiple real-world robot manipulation platforms.", "interface": "RGB images and language map to robot actions.", "limits": "New hardware and action spaces need adaptation. Success rates from different platforms and tasks should not be merged into one capability score.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v270/kim25c.html", "code": "https://github.com/openvla/openvla", "project": "https://openvla.github.io/", "model": "https://huggingface.co/openvla", "pdf": "https://raw.githubusercontent.com/mlresearch/v270/main/assets/kim25c/kim25c.pdf" }, "rank": 2, "publicationNote": "The conference took place in 2024; PMLR proceedings were published in 2025.", "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v270/kim25c.html", "checkedAt": "2026-09-14", "authors": [ "Moo Jin Kim", "Karl Pertsch", "Siddharth Karamcheti", "Ted Xiao", "Ashwin Balakrishna", "Suraj Nair", "Rafael Rafailov", "Ethan P Foster", "Pannag R Sanketi", "Quan Vuong", "Thomas Kollar", "Benjamin Burchfiel", "Russ Tedrake", "Dorsa Sadigh", "Sergey Levine", "Percy Liang", "Chelsea Finn" ], "date": "2025-01-12", "dateLabel": "Publisher record date", "bibtex": "@misc{openvla,\n title = {OpenVLA: An Open-Source Vision-Language-Action Model},\n author = {Moo Jin Kim and Karl Pertsch and Siddharth Karamcheti and Ted Xiao and Ashwin Balakrishna and Suraj Nair and Rafael Rafailov and Ethan P Foster and Pannag R Sanketi and Quan Vuong and Thomas Kollar and Benjamin Burchfiel and Russ Tedrake and Dorsa Sadigh and Sergey Levine and Percy Liang and Chelsea Finn},\n year = {2025},\n url = {https://proceedings.mlr.press/v270/kim25c.html}\n}", "preview": "assets/papers/openvla.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v270/main/assets/kim25c/kim25c.pdf" }, "metrics": { "citations": { "value": 60, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4399695759", "updatedAt": "2026-10-19", "recordTitle": "OpenVLA: An Open-Source Vision-Language-Action Model" }, "github": { "value": 7171, "source": "GitHub", "sourceUrl": "https://github.com/openvla/openvla/stargazers", "repository": "openvla/openvla", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "diffusion-policy", "institutions": [ { "name": "Columbia University", "url": "https://www.columbia.edu/", "mark": "CU" } ], "name": "Diffusion Policy", "title": "Diffusion Policy: Visuomotor Policy Learning via Action Diffusion", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "RSS 2023", "summary": "Models robot action sequences as conditional denoising, generating multimodal continuous actions from visual observations.", "tags": [ "Diffusion models", "Imitation learning" ], "environment": "Simulated and real-world robot manipulation tasks.", "interface": "Visual observations map to continuous action sequences.", "limits": "Denoising steps, control frequency, and demonstration distribution affect performance. Action generation and environment-frame generation are distinct problems.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://diffusion-policy.cs.columbia.edu/", "code": "https://github.com/real-stanford/diffusion_policy", "pdf": "https://arxiv.org/pdf/2303.04137" }, "rank": 4, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2303.04137", "checkedAt": "2026-09-14", "authors": [ "Cheng Chi", "Zhenjia Xu", "Siyuan Feng", "Eric Cousineau", "Yilun Du", "Benjamin Burchfiel", "Russ Tedrake", "Shuran Song" ], "date": "2023-03-07", "dateLabel": "First arXiv submission", "arxivId": "2303.04137", "bibtex": "@misc{diffusion-policy,\n title = {Diffusion Policy: Visuomotor Policy Learning via Action Diffusion},\n author = {Cheng Chi and Zhenjia Xu and Siyuan Feng and Eric Cousineau and Yilun Du and Benjamin Burchfiel and Russ Tedrake and Shuran Song},\n year = {2023},\n url = {https://diffusion-policy.cs.columbia.edu/}\n}", "preview": "assets/papers/diffusion-policy.png", "previewSourceUrl": "https://arxiv.org/pdf/2303.04137" }, "metrics": { "citations": { "value": 524, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4385403811", "updatedAt": "2026-10-19", "recordTitle": "Diffusion Policy: Visuomotor Policy Learning via Action Diffusion" }, "github": { "value": 4656, "source": "GitHub", "sourceUrl": "https://github.com/real-stanford/diffusion_policy/stargazers", "repository": "real-stanford/diffusion_policy", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "act", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "ACT / ALOHA", "title": "Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "RSS 2023", "summary": "Learns fine bimanual manipulation from low-cost teleoperation demonstrations by predicting action chunks with a transformer.", "tags": [ "Imitation learning", "Robot learning" ], "environment": "The ALOHA bimanual platform.", "interface": "Images and joint observations map to action chunks.", "limits": "Requires hardware, teleoperation demonstrations, and task adaptation. Reported success rates do not generalize to arbitrary fine-manipulation tasks.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://roboticsproceedings.org/rss19/p016.html", "project": "https://tonyzhaozh.github.io/aloha/", "code": "https://github.com/tonyzhaozh/act", "pdf": "https://www.roboticsproceedings.org/rss19/p016.pdf" }, "rank": 12, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://roboticsproceedings.org/rss19/p016.html", "checkedAt": "2026-09-14", "authors": [ "Tony Z. Zhao", "Vikash Kumar", "Sergey Levine", "Chelsea Finn" ], "date": "2023-07-10", "dateLabel": "Publisher record date", "bibtex": "@misc{act,\n title = {Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware},\n author = {Tony Z. Zhao and Vikash Kumar and Sergey Levine and Chelsea Finn},\n year = {2023},\n url = {https://roboticsproceedings.org/rss19/p016.html}\n}", "preview": "assets/papers/act.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss19/p016.pdf" }, "metrics": { "citations": { "value": 656, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4385430674", "updatedAt": "2026-10-19", "recordTitle": "Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware" }, "github": { "value": 2281, "source": "GitHub", "sourceUrl": "https://github.com/tonyzhaozh/act/stargazers", "repository": "tonyzhaozh/act", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "octo", "institutions": [ { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" } ], "name": "Octo", "title": "Octo: An Open-Source Generalist Robot Policy", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "RSS 2024", "summary": "Pretrains a generalist policy on demonstrations from multiple robots, with language or goal-image conditioning and adaptation to new inputs and action spaces.", "tags": [ "Robot learning", "Vision-language-action" ], "environment": "Multiple robot platforms.", "interface": "Language or goal images condition a diffusion action policy.", "limits": "New platforms generally still require demonstrations and fine-tuning; an open pretrained policy is not plug-and-play for arbitrary robots.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://www.roboticsproceedings.org/rss20/p090.html", "project": "https://octo-models.github.io/", "code": "https://github.com/octo-models/octo", "model": "https://huggingface.co/rail-berkeley", "pdf": "https://www.roboticsproceedings.org/rss20/p090.pdf" }, "rank": 50, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.roboticsproceedings.org/rss20/p090.html", "checkedAt": "2026-09-14", "authors": [ "Dibya Ghosh", "Homer Rich Walke", "Karl Pertsch", "Kevin Black", "Oier Mees", "Sudeep Dasari", "Joey Hejna", "Tobias Kreiman", "Charles Xu", "Jianlan Luo", "You Liang Tan", "Lawrence Yunliang Chen", "Quan Vuong", "Ted Xiao", "Pannag R Sanketi", "Dorsa Sadigh", "Chelsea Finn", "Sergey Levine" ], "date": "2024-07-15", "dateLabel": "Publisher record date", "bibtex": "@misc{octo,\n title = {Octo: An Open-Source Generalist Robot Policy},\n author = {Dibya Ghosh and Homer Rich Walke and Karl Pertsch and Kevin Black and Oier Mees and Sudeep Dasari and Joey Hejna and Tobias Kreiman and Charles Xu and Jianlan Luo and You Liang Tan and Lawrence Yunliang Chen and Quan Vuong and Ted Xiao and Pannag R Sanketi and Dorsa Sadigh and Chelsea Finn and Sergey Levine},\n year = {2024},\n url = {https://www.roboticsproceedings.org/rss20/p090.html}\n}", "preview": "assets/papers/octo.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss20/p090.pdf" }, "metrics": { "citations": { "value": 118, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402353985", "updatedAt": "2026-10-19", "recordTitle": "Octo: An Open-Source Generalist Robot Policy" }, "github": { "value": 1807, "source": "GitHub", "sourceUrl": "https://github.com/octo-models/octo/stargazers", "repository": "octo-models/octo", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "rt2", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "RT-2", "title": "RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "CoRL 2023", "summary": "Represents robot actions as tokens and co-trains them with vision-language tasks to study transfer of web knowledge into manipulation.", "tags": [ "Vision-language-action", "Robot learning" ], "environment": "Real-world robot manipulation.", "interface": "Images and language map to action tokens.", "limits": "Assess semantic generalization separately from physical skill. This index does not provide a complete runnable model.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v229/zitkovich23a.html", "pdf": "https://proceedings.mlr.press/v229/zitkovich23a/zitkovich23a.pdf" }, "rank": 51, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v229/zitkovich23a.html", "checkedAt": "2026-09-14", "authors": [ "Brianna Zitkovich", "Tianhe Yu", "Sichun Xu", "Peng Xu", "Ted Xiao", "Fei Xia", "Jialin Wu", "Paul Wohlhart", "Stefan Welker", "Ayzaan Wahid", "Quan Vuong", "Vincent Vanhoucke", "Huong Tran", "Radu Soricut", "Anikait Singh", "Jaspiar Singh", "Pierre Sermanet", "Pannag R. Sanketi", "Grecia Salazar", "Michael S. Ryoo", "Krista Reymann", "Kanishka Rao", "Karl Pertsch", "Igor Mordatch", "Henryk Michalewski", "Yao Lu", "Sergey Levine", "Lisa Lee", "Tsang-Wei Edward Lee", "Isabel Leal", "Yuheng Kuang", "Dmitry Kalashnikov", "Ryan Julian", "Nikhil J. Joshi", "Alex Irpan", "Brian Ichter", "Jasmine Hsu", "Alexander Herzog", "Karol Hausman", "Keerthana Gopalakrishnan", "Chuyuan Fu", "Pete Florence", "Chelsea Finn", "Kumar Avinava Dubey", "Danny Driess", "Tianli Ding", "Krzysztof Marcin Choromanski", "Xi Chen", "Yevgen Chebotar", "Justice Carbajal", "Noah Brown", "Anthony Brohan", "Montserrat Gonzalez Arenas", "Kehang Han" ], "date": "2023-12-02", "dateLabel": "Publisher record date", "bibtex": "@misc{rt2,\n title = {RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control},\n author = {Brianna Zitkovich and Tianhe Yu and Sichun Xu and Peng Xu and Ted Xiao and Fei Xia and Jialin Wu and Paul Wohlhart and Stefan Welker and Ayzaan Wahid and Quan Vuong and Vincent Vanhoucke and Huong Tran and Radu Soricut and Anikait Singh and Jaspiar Singh and Pierre Sermanet and Pannag R. Sanketi and Grecia Salazar and Michael S. Ryoo and Krista Reymann and Kanishka Rao and Karl Pertsch and Igor Mordatch and Henryk Michalewski and Yao Lu and Sergey Levine and Lisa Lee and Tsang-Wei Edward Lee and Isabel Leal and Yuheng Kuang and Dmitry Kalashnikov and Ryan Julian and Nikhil J. Joshi and Alex Irpan and Brian Ichter and Jasmine Hsu and Alexander Herzog and Karol Hausman and Keerthana Gopalakrishnan and Chuyuan Fu and Pete Florence and Chelsea Finn and Kumar Avinava Dubey and Danny Driess and Tianli Ding and Krzysztof Marcin Choromanski and Xi Chen and Yevgen Chebotar and Justice Carbajal and Noah Brown and Anthony Brohan and Montserrat Gonzalez Arenas and Kehang Han},\n year = {2023},\n url = {https://proceedings.mlr.press/v229/zitkovich23a.html}\n}", "preview": "assets/papers/rt2.png", "previewSourceUrl": "https://proceedings.mlr.press/v229/zitkovich23a/zitkovich23a.pdf" }, "metrics": { "citations": { "value": 275, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4385473486", "updatedAt": "2026-10-19", "recordTitle": "RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "voxposer", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "VoxPoser", "title": "VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "CoRL 2023", "summary": "Uses language models to compose perception outputs and 3D value maps into objectives and constraints for robot motion planning.", "tags": [ "Programs & tools", "Planning & memory" ], "environment": "RLBench and real-world robot manipulation.", "interface": "Language and perception produce 3D value maps for motion planning.", "limits": "The complete real-world perception pipeline and calibration require separate setup. A generated plan does not guarantee executable motion.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v229/huang23b.html", "project": "https://voxposer.github.io/", "code": "https://github.com/huangwl18/VoxPoser", "pdf": "https://proceedings.mlr.press/v229/huang23b/huang23b.pdf" }, "rank": 52, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v229/huang23b.html", "checkedAt": "2026-09-14", "authors": [ "Wenlong Huang", "Chen Wang", "Ruohan Zhang", "Yunzhu Li", "Jiajun Wu", "Li Fei-Fei" ], "date": "2023-12-02", "dateLabel": "Publisher record date", "bibtex": "@misc{voxposer,\n title = {VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models},\n author = {Wenlong Huang and Chen Wang and Ruohan Zhang and Yunzhu Li and Jiajun Wu and Li Fei-Fei},\n year = {2023},\n url = {https://proceedings.mlr.press/v229/huang23b.html}\n}", "preview": "assets/papers/voxposer.png", "previewSourceUrl": "https://proceedings.mlr.press/v229/huang23b/huang23b.pdf" }, "metrics": { "citations": { "value": 81, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4384264726", "updatedAt": "2026-10-19", "recordTitle": "VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models" }, "github": { "value": 836, "source": "GitHub", "sourceUrl": "https://github.com/huangwl18/VoxPoser/stargazers", "repository": "huangwl18/VoxPoser", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "rekep", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "ReKep", "title": "ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "CoRL 2024", "summary": "Translates instructions into spatial and temporal constraints between 3D keypoints for optimization-based, multi-stage robot manipulation.", "tags": [ "Programs & tools", "Planning & memory" ], "environment": "Single-arm, bimanual, and mobile manipulation.", "interface": "Language and RGB-D observations produce relational keypoint constraints.", "limits": "Keypoint estimation, geometric calibration, and low-level planning jointly determine execution quality.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v270/huang25g.html", "project": "https://rekep-robot.github.io/", "code": "https://github.com/huangwl18/ReKep", "pdf": "https://raw.githubusercontent.com/mlresearch/v270/main/assets/huang25g/huang25g.pdf" }, "rank": 53, "publicationNote": "Conference year: 2024. PMLR publication year: 2025.", "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v270/huang25g.html", "checkedAt": "2026-09-14", "authors": [ "Wenlong Huang", "Chen Wang", "Yunzhu Li", "Ruohan Zhang", "Li Fei-Fei" ], "date": "2025-01-12", "dateLabel": "Publisher record date", "bibtex": "@misc{rekep,\n title = {ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation},\n author = {Wenlong Huang and Chen Wang and Yunzhu Li and Ruohan Zhang and Li Fei-Fei},\n year = {2025},\n url = {https://proceedings.mlr.press/v270/huang25g.html}\n}", "preview": "assets/papers/rekep.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v270/main/assets/huang25g/huang25g.pdf" }, "metrics": { "citations": { "value": 12, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402955063", "updatedAt": "2026-10-19", "recordTitle": "ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulation" }, "github": { "value": 989, "source": "GitHub", "sourceUrl": "https://github.com/huangwl18/ReKep/stargazers", "repository": "huangwl18/ReKep", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "hirobot", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "Hi Robot", "title": "Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "ICML 2025", "summary": "Uses a high-level vision-language model to interpret complex instructions and feedback, while a low-level vision-language-action model executes subtasks.", "tags": [ "Vision-language-action", "Planning & memory" ], "environment": "Single-arm, bimanual, and mobile robots.", "interface": "High-level language subgoals guide low-level actions.", "limits": "Task decomposition, subgoal transitions, and low-level skill coverage jointly affect long-horizon success.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v267/shi25d.html", "pdf": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/shi25d/shi25d.pdf" }, "rank": 8, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v267/shi25d.html", "checkedAt": "2026-09-14", "authors": [ "Lucy Xiaoyang Shi", "Brian Ichter", "Michael Robert Equi", "Liyiming Ke", "Karl Pertsch", "Quan Vuong", "James Tanner", "Anna Walling", "Haohuan Wang", "Niccolo Fusai", "Adrian Li-Bell", "Danny Driess", "Lachy Groom", "Sergey Levine", "Chelsea Finn" ], "date": "2025-10-06", "dateLabel": "Publisher record date", "bibtex": "@misc{hirobot,\n title = {Hi Robot: Open-Ended Instruction Following with Hierarchical Vision-Language-Action Models},\n author = {Lucy Xiaoyang Shi and Brian Ichter and Michael Robert Equi and Liyiming Ke and Karl Pertsch and Quan Vuong and James Tanner and Anna Walling and Haohuan Wang and Niccolo Fusai and Adrian Li-Bell and Danny Driess and Lachy Groom and Sergey Levine and Chelsea Finn},\n year = {2025},\n url = {https://proceedings.mlr.press/v267/shi25d.html}\n}", "preview": "assets/papers/hirobot.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/shi25d/shi25d.pdf" }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "libero", "name": "LIBERO", "title": "LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning", "domains": [ "embodied" ], "kinds": [ "papers", "benchmarks" ], "year": 2023, "venue": "NeurIPS 2023 D&B", "summary": "Evaluates lifelong robot learning and knowledge transfer across changes in spatial relations, objects, goals, and task combinations.", "tags": [ "Benchmarks", "Continual learning" ], "environment": "Simulated robot manipulation.", "interface": "Language goals with simulated observations and actions.", "limits": "Specify the task suite, demonstrations, task order, and evaluation protocol. Simulation results do not establish real-world deployment performance.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/8c3c666820ea055a77726d66fc7d447f-Abstract-Datasets_and_Benchmarks.html", "code": "https://github.com/Lifelong-Robot-Learning/LIBERO", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/8c3c666820ea055a77726d66fc7d447f-Paper-Datasets_and_Benchmarks.pdf" }, "rank": 16, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/8c3c666820ea055a77726d66fc7d447f-Abstract-Datasets_and_Benchmarks.html", "checkedAt": "2026-09-14", "authors": [ "Bo Liu", "Yifeng Zhu", "Chongkai Gao", "Yihao Feng", "Qiang Liu", "Yuke Zhu", "Peter Stone" ], "date": "2023-12-15", "dateLabel": "Publisher record date", "doi": "10.52202/075280-1939", "bibtex": "@misc{libero,\n title = {LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning},\n author = {Bo Liu and Yifeng Zhu and Chongkai Gao and Yihao Feng and Qiang Liu and Yuke Zhu and Peter Stone},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/8c3c666820ea055a77726d66fc7d447f-Abstract-Datasets_and_Benchmarks.html},\n doi = {10.52202/075280-1939}\n}", "preview": "assets/papers/libero.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/8c3c666820ea055a77726d66fc7d447f-Paper-Datasets_and_Benchmarks.pdf" }, "metrics": { "citations": { "value": 7, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4379919389", "updatedAt": "2026-10-12", "recordTitle": "LIBERO: Benchmarking Knowledge Transfer for Lifelong Robot Learning" }, "github": { "value": 2445, "source": "GitHub", "sourceUrl": "https://github.com/Lifelong-Robot-Learning/LIBERO/stargazers", "repository": "Lifelong-Robot-Learning/LIBERO", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "embodiedbench", "name": "EmbodiedBench", "title": "EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents", "domains": [ "embodied" ], "kinds": [ "papers", "benchmarks" ], "year": 2025, "venue": "ICML 2025", "summary": "Systematically evaluates multimodal language models on decision-making in visually grounded embodied tasks.", "tags": [ "Benchmarks", "Language & action" ], "environment": "Multiple embodied simulation tasks and environments.", "interface": "Visual observations map to environment-defined actions.", "limits": "Task types, action abstraction, and environment settings differ; aggregate success rates alone do not support direct comparisons.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v267/yang25f.html", "project": "https://embodiedbench.github.io/", "code": "https://github.com/EmbodiedBench/EmbodiedBench", "data": "https://huggingface.co/EmbodiedBench", "pdf": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/yang25f/yang25f.pdf" }, "rank": 56, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v267/yang25f.html", "checkedAt": "2026-09-14", "authors": [ "Rui Yang", "Hanyang Chen", "Junyu Zhang", "Mark Zhao", "Cheng Qian", "Kangrui Wang", "Qineng Wang", "Teja Venkat Koripella", "Marziyeh Movahedi", "Manling Li", "Heng Ji", "Huan Zhang", "Tong Zhang" ], "date": "2025-10-06", "dateLabel": "Publisher record date", "bibtex": "@misc{embodiedbench,\n title = {EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents},\n author = {Rui Yang and Hanyang Chen and Junyu Zhang and Mark Zhao and Cheng Qian and Kangrui Wang and Qineng Wang and Teja Venkat Koripella and Marziyeh Movahedi and Manling Li and Heng Ji and Huan Zhang and Tong Zhang},\n year = {2025},\n url = {https://proceedings.mlr.press/v267/yang25f.html}\n}", "preview": "assets/papers/embodiedbench.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v267/main/assets/yang25f/yang25f.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4407590542", "updatedAt": "2026-10-19", "recordTitle": "EmbodiedBench: Comprehensive Benchmarking Multi-modal Large Language Models for Vision-Driven Embodied Agents" }, "github": { "value": 353, "source": "GitHub", "sourceUrl": "https://github.com/EmbodiedBench/EmbodiedBench/stargazers", "repository": "EmbodiedBench/EmbodiedBench", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "consistency-policy", "name": "Consistency Policy", "title": "Consistency Policy: Accelerated Visuomotor Policies via Consistency Distillation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "RSS 2024", "summary": "Distills diffusion policies into fewer action-generation steps to study visual-control latency under limited compute.", "tags": [ "Diffusion models", "Robot learning" ], "environment": "Simulated and real-world manipulation tasks.", "interface": "Visual observations map to few-step action generation.", "limits": "Speedups depend on hardware, baseline sampling settings, and task. Generation throughput is not end-to-end control frequency.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://www.roboticsproceedings.org/rss20/p071.html", "pdf": "https://www.roboticsproceedings.org/rss20/p071.pdf" }, "rank": 57, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.roboticsproceedings.org/rss20/p071.html", "checkedAt": "2026-09-14", "authors": [ "Aaditya Prasad", "Kevin Lin", "Jimmy Wu", "Linqi Zhou", "Jeannette Bohg" ], "date": "2024-07-15", "dateLabel": "Publisher record date", "bibtex": "@misc{consistency-policy,\n title = {Consistency Policy: Accelerated Visuomotor Policies via Consistency Distillation},\n author = {Aaditya Prasad and Kevin Lin and Jimmy Wu and Linqi Zhou and Jeannette Bohg},\n year = {2024},\n url = {https://www.roboticsproceedings.org/rss20/p071.html}\n}", "preview": "assets/papers/consistency-policy.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss20/p071.pdf" }, "metrics": { "citations": { "value": 54, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402354007", "updatedAt": "2026-10-19", "recordTitle": "Consistency Policy: Accelerated Visuomotor Policies via Consistency Distillation" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "baku", "name": "BAKU", "title": "BAKU: An Efficient Transformer for Multi-Task Policy Learning", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "NeurIPS 2024", "summary": "Uses transformers for multitask robot policies, studying efficient architectures and action-prediction designs.", "tags": [ "Robot learning", "Imitation learning" ], "environment": "LIBERO, Meta-World, and control benchmarks.", "interface": "Multimodal observations feed a multitask action policy.", "limits": "Report benchmarks and training budgets separately. Experiments have not been independently reproduced for this index.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/ff887781480973bd3cb6026feb378d1e-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2024/file/ff887781480973bd3cb6026feb378d1e-Paper-Conference.pdf" }, "rank": 58, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2024/hash/ff887781480973bd3cb6026feb378d1e-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Siddhant Haldar", "Zhuoran Peng", "Lerrel Pinto" ], "date": "2024-12-16", "dateLabel": "Publisher record date", "doi": "10.52202/079017-4484", "bibtex": "@misc{baku,\n title = {BAKU: An Efficient Transformer for Multi-Task Policy Learning},\n author = {Siddhant Haldar and Zhuoran Peng and Lerrel Pinto},\n year = {2024},\n url = {https://proceedings.neurips.cc/paper_files/paper/2024/hash/ff887781480973bd3cb6026feb378d1e-Abstract-Conference.html},\n doi = {10.52202/079017-4484}\n}", "preview": "assets/papers/baku.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2024/file/ff887781480973bd3cb6026feb378d1e-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4399598096", "updatedAt": "2026-10-05", "recordTitle": "BAKU: An Efficient Transformer for Multi-Task Policy Learning" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "vla-cache", "name": "VLA-Cache", "title": "VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "NeurIPS 2025", "summary": "Reuses visual-token computation to accelerate vision-language-action models, examining the tradeoff between latency and manipulation success.", "tags": [ "Vision-language-action", "Robot learning" ], "environment": "LIBERO, SIMPLER, and real robots.", "interface": "Token caching within vision-language-action models.", "limits": "Speed depends on the model, hardware, and scene changes. Acceleration factors measured on different devices are not directly comparable.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/f062da1973ac9ac61fc6d44dd7fa309f-Abstract-Conference.html", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2025/file/f062da1973ac9ac61fc6d44dd7fa309f-Paper-Conference.pdf" }, "rank": 59, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2025/hash/f062da1973ac9ac61fc6d44dd7fa309f-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Siyu Xu", "Yunke Wang", "Chenghao Xia", "Dihao Zhu", "Tao Huang", "Chang Xu" ], "date": "2026-04-23", "dateLabel": "Publisher record date", "doi": "10.52202/085713-5484", "bibtex": "@misc{vla-cache,\n title = {VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching},\n author = {Siyu Xu and Yunke Wang and Chenghao Xia and Dihao Zhu and Tao Huang and Chang Xu},\n year = {2026},\n url = {https://proceedings.neurips.cc/paper_files/paper/2025/hash/f062da1973ac9ac61fc6d44dd7fa309f-Abstract-Conference.html},\n doi = {10.52202/085713-5484}\n}", "preview": "assets/papers/vla-cache.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2025/file/f062da1973ac9ac61fc6d44dd7fa309f-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4407185791", "updatedAt": "2026-09-21", "recordTitle": "VLA-Cache: Efficient Vision-Language-Action Manipulation via Adaptive Token Caching" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "robochemist", "name": "RoboChemist", "title": "RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "CoRL 2025", "summary": "Combines high-level vision-language planning and low-level action models for multi-stage laboratory tasks, with feedback and procedural constraints.", "tags": [ "Planning & memory", "Safety & verification" ], "environment": "Robotic chemistry experiments.", "interface": "Vision-language planning and monitoring with VLA execution.", "limits": "Evidence of protocol adherence in a research setting does not constitute general laboratory safety certification.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v305/zhang25i.html", "pdf": "https://raw.githubusercontent.com/mlresearch/v305/main/assets/zhang25i/zhang25i.pdf" }, "rank": 60, "selectionTrack": "peer-reviewed", "selectionReason": "Conference or journal paper within the five-year window; official proceedings and author sources are prioritized.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v305/zhang25i.html", "checkedAt": "2026-09-14", "authors": [ "Zongzheng Zhang", "Chenghao Yue", "Haobo Xu", "Minwen Liao", "Xianglin Qi", "Huan-ang Gao", "Ziwei Wang", "Hao Zhao" ], "date": "2025-10-07", "dateLabel": "Publisher record date", "bibtex": "@misc{robochemist,\n title = {RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation},\n author = {Zongzheng Zhang and Chenghao Yue and Haobo Xu and Minwen Liao and Xianglin Qi and Huan-ang Gao and Ziwei Wang and Hao Zhao},\n year = {2025},\n url = {https://proceedings.mlr.press/v305/zhang25i.html}\n}", "preview": "assets/papers/robochemist.png", "previewSourceUrl": "https://raw.githubusercontent.com/mlresearch/v305/main/assets/zhang25i/zhang25i.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4417071139", "updatedAt": "2026-09-21", "recordTitle": "RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "sima2", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "SIMA 2", "title": "SIMA 2: A Generalist Embodied Agent for Virtual Worlds", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "arXiv · 2025.12", "summary": "Connects Gemini reasoning and interaction to multiple 3D games, studying instruction following, cross-environment generalization, and self-improvement.", "tags": [ "Language & action", "Exploration & skills" ], "environment": "Multiple 3D virtual worlds.", "interface": "Vision and language map to keyboard and mouse interaction.", "limits": "The evidence concerns virtual environments. It does not establish arbitrary long-horizon competence or solved real-world robot control.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2512.04797", "project": "https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/", "pdf": "https://arxiv.org/pdf/2512.04797" }, "rank": 6, "publicationDate": "2025-12-04", "selectionTrack": "recent-arxiv", "selectionReason": "Editorial selection: official cross-game evaluations, interactive reasoning, and self-improvement results provide substantial system-level evidence for generalist game agents. Inclusion is not based solely on institutional reputation or popularity.", "influenceSources": [ "https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/" ], "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2512.04797", "checkedAt": "2026-09-14", "authors": [ "SIMA team", "Adrian Bolton", "Alexander Lerchner", "Alexandra Cordell", "Alexandre Moufarek", "Andrew Bolt", "Andrew Lampinen", "Anna Mitenkova", "Arne Olav Hallingstad", "Bojan Vujatovic", "Bonnie Li", "Cong Lu", "Daan Wierstra", "Daniel P. Sawyer", "Daniel Slater", "David Reichert", "Davide Vercelli", "Demis Hassabis", "Drew A. Hudson", "Duncan Williams", "Ed Hirst", "Fabio Pardo", "Felix Hill", "Frederic Besse", "Hannah Openshaw", "Harris Chan", "Hubert Soyer", "Jane X. Wang", "Jeff Clune", "John Agapiou", "John Reid", "Joseph Marino", "Junkyung Kim", "Karol Gregor", "Kaustubh Sridhar", "Kay McKinney", "Laura Kampis", "Lei M. Zhang", "Loic Matthey", "Luyu Wang", "Maria Abi Raad", "Maria Loks-Thompson", "Martin Engelcke", "Matija Kecman", "Matthew Jackson", "Maxime Gazeau", "Ollie Purkiss", "Oscar Knagg", "Peter Stys", "Piermaria Mendolicchio", "Raia Hadsell", "Rosemary Ke", "Ryan Faulkner", "Sarah Chakera", "Satinder Singh Baveja", "Shane Legg", "Sheleem Kashem", "Tayfun Terzi", "Thomas Keck", "Tim Harley", "Tim Scholtes", "Tyson Roberts", "Volodymyr Mnih", "Yulan Liu", "Zhengdong Wang", "Zoubin Ghahramani" ], "date": "2025-12-04", "dateLabel": "First arXiv submission", "arxivId": "2512.04797", "bibtex": "@misc{sima2,\n title = {SIMA 2: A Generalist Embodied Agent for Virtual Worlds},\n author = {{SIMA team} and Adrian Bolton and Alexander Lerchner and Alexandra Cordell and Alexandre Moufarek and Andrew Bolt and Andrew Lampinen and Anna Mitenkova and Arne Olav Hallingstad and Bojan Vujatovic and Bonnie Li and Cong Lu and Daan Wierstra and Daniel P. Sawyer and Daniel Slater and David Reichert and Davide Vercelli and Demis Hassabis and Drew A. Hudson and Duncan Williams and Ed Hirst and Fabio Pardo and Felix Hill and Frederic Besse and Hannah Openshaw and Harris Chan and Hubert Soyer and Jane X. Wang and Jeff Clune and John Agapiou and John Reid and Joseph Marino and Junkyung Kim and Karol Gregor and Kaustubh Sridhar and Kay McKinney and Laura Kampis and Lei M. Zhang and Loic Matthey and Luyu Wang and Maria Abi Raad and Maria Loks-Thompson and Martin Engelcke and Matija Kecman and Matthew Jackson and Maxime Gazeau and Ollie Purkiss and Oscar Knagg and Peter Stys and Piermaria Mendolicchio and Raia Hadsell and Rosemary Ke and Ryan Faulkner and Sarah Chakera and Satinder Singh Baveja and Shane Legg and Sheleem Kashem and Tayfun Terzi and Thomas Keck and Tim Harley and Tim Scholtes and Tyson Roberts and Volodymyr Mnih and Yulan Liu and Zhengdong Wang and Zoubin Ghahramani},\n year = {2025},\n url = {https://arxiv.org/abs/2512.04797}\n}", "preview": "assets/papers/sima2.png", "previewSourceUrl": "https://arxiv.org/pdf/2512.04797" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4417086995", "updatedAt": "2026-09-28", "recordTitle": "SIMA 2: A Generalist Embodied Agent for Virtual Worlds" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "pi0-paper", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" } ], "name": "π0", "title": "π0: A Vision-Language-Action Flow Model for General Robot Control", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "RSS 2025", "summary": "Introduces a generalist VLA that augments a pretrained vision-language model with a flow-matching action expert for dexterous, high-frequency robot control.", "tags": [ "Vision-language-action", "Flow matching" ], "environment": "More than 10,000 hours of cross-embodiment robot data spanning single-arm, dual-arm, and mobile manipulation tasks.", "interface": "Images, robot state, and language prompts map to continuous action chunks at control rates up to 50 Hz.", "limits": "The evaluation uses the team's robot platforms and data mixture; out-of-the-box and task-specific post-training results should be distinguished.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://www.roboticsproceedings.org/rss21/p010.html", "project": "https://www.pi.website/blog/pi0", "pdf": "https://arxiv.org/pdf/2410.24164", "code": "https://github.com/Physical-Intelligence/openpi" }, "rank": 2, "selectionTrack": "peer-reviewed", "selectionReason": "Published at a leading robotics or machine-learning venue within the five-year window; official paper and project sources verified.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2410.24164", "checkedAt": "2026-09-15", "authors": [ "Kevin Black", "Noah Brown", "Danny Driess", "Adnan Esmail", "Michael Equi", "Chelsea Finn", "Niccolo Fusai", "Lachy Groom", "Karol Hausman", "Brian Ichter", "Szymon Jakubczak", "Tim Jones", "Liyiming Ke", "Sergey Levine", "Adrian Li-Bell", "Mohith Mothukuri", "Suraj Nair", "Karl Pertsch", "Lucy Xiaoyang Shi", "Laura Smith", "James Tanner", "Quan Vuong", "Anna Walling", "Haohuan Wang", "Ury Zhilinsky" ], "date": "2024-10-31", "dateLabel": "First arXiv submission", "arxivId": "2410.24164", "bibtex": "@inproceedings{pi0-paper,\n title = {π0: A Vision-Language-Action Flow Model for General Robot Control},\n author = {Kevin Black and Noah Brown and Danny Driess and Adnan Esmail and Michael Equi and Chelsea Finn and Niccolo Fusai and Lachy Groom and Karol Hausman and Brian Ichter and Szymon Jakubczak and Tim Jones and Liyiming Ke and Sergey Levine and Adrian Li-Bell and Mohith Mothukuri and Suraj Nair and Karl Pertsch and Lucy Xiaoyang Shi and Laura Smith and James Tanner and Quan Vuong and Anna Walling and Haohuan Wang and Ury Zhilinsky},\n year = {2025},\n booktitle = {RSS},\n url = {https://arxiv.org/abs/2410.24164}\n}", "preview": "assets/papers/pi0-paper.png", "previewSourceUrl": "https://arxiv.org/pdf/2410.24164", "doi": "10.15607/RSS.2025.XXI.010" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "metrics": { "citations": { "value": 229, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4414079054", "updatedAt": "2026-09-15", "recordTitle": "π0: A Vision-Language-Action Flow Model for General Robot Control" }, "github": { "value": 14293, "source": "GitHub", "sourceUrl": "https://github.com/Physical-Intelligence/openpi/stargazers", "repository": "Physical-Intelligence/openpi", "updatedAt": "2026-10-19" } } }, { "id": "pi0-fast", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" }, { "name": "UC Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" }, { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "π0-FAST", "title": "FAST: Efficient Action Tokenization for Vision-Language-Action Models", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "RSS 2025", "summary": "Compresses continuous action chunks with a discrete cosine transform and byte-pair encoding, enabling efficient autoregressive VLA training and the π0-FAST policy.", "tags": [ "Vision-language-action", "Action tokenization" ], "environment": "Dexterous manipulation and DROID generalization experiments, including policies trained on roughly 10,000 hours of robot data.", "interface": "Continuous robot trajectories are converted into compact frequency-space tokens for autoregressive prediction.", "limits": "FAST improves token efficiency, but performance still depends on the policy backbone, data mixture, robot interface, and deployment calibration.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://www.roboticsproceedings.org/rss21/p012.html", "project": "https://www.pi.website/research/fast", "pdf": "https://arxiv.org/pdf/2501.09747", "code": "https://github.com/Physical-Intelligence/openpi" }, "rank": 3, "selectionTrack": "peer-reviewed", "selectionReason": "Published at a leading robotics or machine-learning venue within the five-year window; official paper and project sources verified.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2501.09747", "checkedAt": "2026-09-15", "authors": [ "Karl Pertsch", "Kyle Stachowicz", "Brian Ichter", "Danny Driess", "Suraj Nair", "Quan Vuong", "Oier Mees", "Chelsea Finn", "Sergey Levine" ], "date": "2025-01-16", "dateLabel": "First arXiv submission", "arxivId": "2501.09747", "bibtex": "@inproceedings{pi0-fast,\n title = {FAST: Efficient Action Tokenization for Vision-Language-Action Models},\n author = {Karl Pertsch and Kyle Stachowicz and Brian Ichter and Danny Driess and Suraj Nair and Quan Vuong and Oier Mees and Chelsea Finn and Sergey Levine},\n year = {2025},\n booktitle = {RSS},\n url = {https://arxiv.org/abs/2501.09747}\n}", "preview": "assets/papers/pi0-fast.png", "previewSourceUrl": "https://arxiv.org/pdf/2501.09747", "doi": "10.15607/RSS.2025.XXI.012" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "metrics": { "citations": { "value": 70, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4414051084", "updatedAt": "2026-10-19", "recordTitle": "FAST: Efficient Action Tokenization for Vision-Language-Action Models" }, "github": { "value": 14293, "source": "GitHub", "sourceUrl": "https://github.com/Physical-Intelligence/openpi/stargazers", "repository": "Physical-Intelligence/openpi", "updatedAt": "2026-10-19" } } }, { "id": "pi05", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" } ], "name": "π0.5", "title": "π0.5: a Vision-Language-Action Model with Open-World Generalization", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "CoRL 2025", "summary": "Extends π0 with heterogeneous co-training across robot actions, semantic subtasks, object grounding, language instructions, and web data for open-world generalization.", "tags": [ "Vision-language-action", "Open-world generalization" ], "environment": "Long-horizon mobile manipulation in previously unseen homes, including kitchen and bedroom cleanup.", "interface": "A high-level policy predicts semantic subtasks while a low-level action expert controls the robot from multimodal observations and language.", "limits": "Generalization is evaluated on the reported homes and task families; long-horizon demonstrations still include failures and do not establish unrestricted household autonomy.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://openreview.net/forum?id=vlhoswksBO", "project": "https://www.pi.website/blog/pi05", "pdf": "https://arxiv.org/pdf/2504.16054", "code": "https://github.com/Physical-Intelligence/openpi" }, "rank": 4, "selectionTrack": "peer-reviewed", "selectionReason": "Published at a leading robotics or machine-learning venue within the five-year window; official paper and project sources verified.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2504.16054", "checkedAt": "2026-09-15", "authors": [ "Physical Intelligence", "Kevin Black", "Noah Brown", "James Darpinian", "Karan Dhabalia", "Danny Driess", "Adnan Esmail", "Michael Equi", "Chelsea Finn", "Niccolo Fusai", "Manuel Y. Galliker", "Dibya Ghosh", "Lachy Groom", "Karol Hausman", "Brian Ichter", "Szymon Jakubczak", "Tim Jones", "Liyiming Ke", "Devin LeBlanc", "Sergey Levine", "Adrian Li-Bell", "Mohith Mothukuri", "Suraj Nair", "Karl Pertsch", "Allen Z. Ren", "Lucy Xiaoyang Shi", "Laura Smith", "Jost Tobias Springenberg", "Kyle Stachowicz", "James Tanner", "Quan Vuong", "Homer Walke", "Anna Walling", "Haohuan Wang", "Lili Yu", "Ury Zhilinsky" ], "date": "2025-04-22", "dateLabel": "First arXiv submission", "arxivId": "2504.16054", "bibtex": "@inproceedings{pi05,\n title = {π0.5: a Vision-Language-Action Model with Open-World Generalization},\n author = {Physical Intelligence and Kevin Black and Noah Brown and James Darpinian and Karan Dhabalia and Danny Driess and Adnan Esmail and Michael Equi and Chelsea Finn and Niccolo Fusai and Manuel Y. Galliker and Dibya Ghosh and Lachy Groom and Karol Hausman and Brian Ichter and Szymon Jakubczak and Tim Jones and Liyiming Ke and Devin LeBlanc and Sergey Levine and Adrian Li-Bell and Mohith Mothukuri and Suraj Nair and Karl Pertsch and Allen Z. Ren and Lucy Xiaoyang Shi and Laura Smith and Jost Tobias Springenberg and Kyle Stachowicz and James Tanner and Quan Vuong and Homer Walke and Anna Walling and Haohuan Wang and Lili Yu and Ury Zhilinsky},\n year = {2025},\n booktitle = {CoRL},\n url = {https://arxiv.org/abs/2504.16054}\n}", "preview": "assets/papers/pi05.png", "previewSourceUrl": "https://arxiv.org/pdf/2504.16054" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "metrics": { "citations": { "value": 3, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4414674749", "updatedAt": "2026-09-28", "recordTitle": "$π_{0.5}$: a Vision-Language-Action Model with Open-World Generalization" }, "github": { "value": 14293, "source": "GitHub", "sourceUrl": "https://github.com/Physical-Intelligence/openpi/stargazers", "repository": "Physical-Intelligence/openpi", "updatedAt": "2026-10-19" } } }, { "id": "pi05-ki", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" } ], "name": "π0.5 + KI", "title": "Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "NeurIPS 2025", "summary": "Separates continuous-action gradients from the pretrained VLM backbone while using FAST tokens and vision-language co-training to preserve semantic knowledge and speed up learning.", "tags": [ "Vision-language-action", "Knowledge insulation" ], "environment": "Real-robot mobile and tabletop manipulation plus the DROID and LIBERO benchmarks.", "interface": "The VLM backbone learns from discrete action and vision-language objectives while a separate flow-matching action expert produces continuous actions.", "limits": "Knowledge insulation addresses gradient interference in the studied architecture; it does not remove the need for diverse data or guarantee transfer to every robot.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://papers.nips.cc/paper_files/paper/2025/hash/94e936034d12bcd04834ec2773f02aff-Abstract-Conference.html", "project": "https://www.pi.website/research/knowledge_insulation", "pdf": "https://arxiv.org/pdf/2505.23705", "code": "https://github.com/Physical-Intelligence/openpi" }, "rank": 5, "selectionTrack": "peer-reviewed", "selectionReason": "Published at a leading robotics or machine-learning venue within the five-year window; official paper and project sources verified.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2505.23705", "checkedAt": "2026-09-15", "authors": [ "Danny Driess", "Jost Tobias Springenberg", "Brian Ichter", "Lili Yu", "Adrian Li-Bell", "Karl Pertsch", "Allen Z. Ren", "Homer Walke", "Quan Vuong", "Lucy Xiaoyang Shi", "Sergey Levine" ], "date": "2025-05-29", "dateLabel": "First arXiv submission", "arxivId": "2505.23705", "bibtex": "@inproceedings{pi05-ki,\n title = {Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better},\n author = {Danny Driess and Jost Tobias Springenberg and Brian Ichter and Lili Yu and Adrian Li-Bell and Karl Pertsch and Allen Z. Ren and Homer Walke and Quan Vuong and Lucy Xiaoyang Shi and Sergey Levine},\n year = {2025},\n booktitle = {NeurIPS},\n url = {https://arxiv.org/abs/2505.23705}\n}", "preview": "assets/papers/pi05-ki.png", "previewSourceUrl": "https://arxiv.org/pdf/2505.23705", "doi": "10.52202/085713-3439" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4416613198", "updatedAt": "2026-09-15", "recordTitle": "Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better" }, "github": { "value": 14293, "source": "GitHub", "sourceUrl": "https://github.com/Physical-Intelligence/openpi/stargazers", "repository": "Physical-Intelligence/openpi", "updatedAt": "2026-10-19" } } }, { "id": "pi05-ego", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" }, { "name": "Georgia Institute of Technology", "url": "https://www.gatech.edu/", "mark": "GT" } ], "name": "π0.5 + ego", "title": "Emergence of Human to Robot Transfer in Vision-Language-Action Models", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2026, "venue": "RSS 2026", "summary": "Shows that human-to-robot transfer emerges with sufficiently diverse VLA pretraining, then adapts π0.5 with egocentric human demonstrations through simple co-training.", "tags": [ "Vision-language-action", "Human-to-robot transfer" ], "environment": "Real-world scene, object, and task generalization using wearable-camera human data and multiple robot embodiments.", "interface": "Egocentric human hand trajectories and robot demonstrations share high-level subtask and low-level action learning objectives.", "limits": "The transfer gains depend on diverse robot pretraining and targeted human data; human videos alone do not directly solve the embodiment gap.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://www.roboticsproceedings.org/rss22/p072.html", "project": "https://www.pi.website/research/human_to_robot", "pdf": "https://arxiv.org/pdf/2512.22414" }, "rank": 6, "selectionTrack": "peer-reviewed", "selectionReason": "Published at a leading robotics or machine-learning venue within the five-year window; official paper and project sources verified.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2512.22414", "checkedAt": "2026-09-15", "authors": [ "Simar Kareer", "Karl Pertsch", "James Darpinian", "Judy Hoffman", "Danfei Xu", "Sergey Levine", "Chelsea Finn", "Suraj Nair" ], "date": "2025-12-27", "dateLabel": "First arXiv submission", "arxivId": "2512.22414", "bibtex": "@inproceedings{pi05-ego,\n title = {Emergence of Human to Robot Transfer in Vision-Language-Action Models},\n author = {Simar Kareer and Karl Pertsch and James Darpinian and Judy Hoffman and Danfei Xu and Sergey Levine and Chelsea Finn and Suraj Nair},\n year = {2026},\n booktitle = {RSS},\n url = {https://arxiv.org/abs/2512.22414}\n}", "preview": "assets/papers/pi05-ego.png", "previewSourceUrl": "https://arxiv.org/pdf/2512.22414", "doi": "10.15607/RSS.2026.XXII.072" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7117736774", "updatedAt": "2026-10-19", "recordTitle": "Emergence of Human to Robot Transfer in Vision-Language-Action Models" } } }, { "id": "pi07", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.pi.website/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" } ], "name": "π0.7", "title": "π0.7: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2026, "venue": "arXiv · 2026.04", "summary": "Uses diverse multimodal context—language, subgoal images, control labels, speed, and quality—to steer a single generalist policy and recombine skills across robots and tasks.", "tags": [ "Vision-language-action", "Compositional generalization" ], "environment": "Dexterous and long-horizon real-world manipulation across multiple robot platforms, homes, objects, and kitchen appliances.", "interface": "Text instructions, multimodal prompts, metadata, visual subgoals, observations, and memory condition a high-level policy and continuous action expert.", "limits": "The evidence comes from the authors' controlled evaluations and demonstrations; emergent generalization remains task- and platform-bounded.", "status": "selected", "evidence": "Official paper, project page, and bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-15", "links": { "paper": "https://arxiv.org/abs/2604.15483", "project": "https://www.pi.website/blog/pi07", "pdf": "https://arxiv.org/pdf/2604.15483" }, "rank": 7, "selectionTrack": "recent-arxiv", "selectionReason": "Editorial selection: the official report presents broad controlled evaluations of steerable, compositional, and cross-embodiment behavior in the newest π-series model. Public demonstrations support significance but are not treated as independent reproduction.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2604.15483", "checkedAt": "2026-09-15", "authors": [ "Physical Intelligence", "Bo Ai", "Ali Amin", "Raichelle Aniceto", "Ashwin Balakrishna", "Greg Balke", "Kevin Black", "George Bokinsky", "Shihao Cao", "Thomas Charbonnier", "Vedant Choudhary", "Foster Collins", "Ken Conley", "Grace Connors", "James Darpinian", "Karan Dhabalia", "Maitrayee Dhaka", "Jared DiCarlo", "Danny Driess", "Michael Equi", "Adnan Esmail", "Yunhao Fang", "Chelsea Finn", "Catherine Glossop", "Thomas Godden", "Ivan Goryachev", "Lachlan Groom", "Haroun Habeeb", "Hunter Hancock", "Karol Hausman", "Gashon Hussein", "Victor Hwang", "Brian Ichter", "Connor Jacobsen", "Szymon Jakubczak", "Rowan Jen", "Tim Jones", "Gregg Kammerer", "Ben Katz", "Liyiming Ke", "Mairbek Khadikov", "Chandra Kuchi", "Marinda Lamb", "Devin LeBlanc", "Brendon LeCount", "Sergey Levine", "Xinyu Li", "Adrian Li-Bell", "Vladislav Lialin", "Zhonglin Liang", "Wallace Lim", "Yao Lu", "Enyu Luo", "Vishnu Mano", "Nandan Marwaha", "Aikys Mongush", "Liam Murphy", "Suraj Nair", "Tyler Patterson", "Karl Pertsch", "Allen Z. Ren", "Gavin Schelske", "Charvi Sharma", "Baifeng Shi", "Lucy Xiaoyang Shi", "Laura Smith", "Jost Tobias Springenberg", "Kyle Stachowicz", "Will Stoeckle", "Jiaming Tang", "Jimmy Tanner", "Shalom Tekeste", "Marcel Torne", "Kyle Vedder", "Quan Vuong", "Anna Walling", "Haohuan Wang", "Jason Wang", "XuDong Wang", "Chris Whalen", "Samuel Whitmore", "Blake Williams", "Charles Xu", "Sukwon Yoo", "Lili Yu", "Wuming Zhang", "Zhuoyang Zhang", "Ury Zhilinsky" ], "date": "2026-04-16", "dateLabel": "First arXiv submission", "arxivId": "2604.15483", "bibtex": "@misc{pi07,\n title = {π0.7: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities},\n author = {Physical Intelligence and Bo Ai and Ali Amin and Raichelle Aniceto and Ashwin Balakrishna and Greg Balke and Kevin Black and George Bokinsky and Shihao Cao and Thomas Charbonnier and Vedant Choudhary and Foster Collins and Ken Conley and Grace Connors and James Darpinian and Karan Dhabalia and Maitrayee Dhaka and Jared DiCarlo and Danny Driess and Michael Equi and Adnan Esmail and Yunhao Fang and Chelsea Finn and Catherine Glossop and Thomas Godden and Ivan Goryachev and Lachlan Groom and Haroun Habeeb and Hunter Hancock and Karol Hausman and Gashon Hussein and Victor Hwang and Brian Ichter and Connor Jacobsen and Szymon Jakubczak and Rowan Jen and Tim Jones and Gregg Kammerer and Ben Katz and Liyiming Ke and Mairbek Khadikov and Chandra Kuchi and Marinda Lamb and Devin LeBlanc and Brendon LeCount and Sergey Levine and Xinyu Li and Adrian Li-Bell and Vladislav Lialin and Zhonglin Liang and Wallace Lim and Yao Lu and Enyu Luo and Vishnu Mano and Nandan Marwaha and Aikys Mongush and Liam Murphy and Suraj Nair and Tyler Patterson and Karl Pertsch and Allen Z. Ren and Gavin Schelske and Charvi Sharma and Baifeng Shi and Lucy Xiaoyang Shi and Laura Smith and Jost Tobias Springenberg and Kyle Stachowicz and Will Stoeckle and Jiaming Tang and Jimmy Tanner and Shalom Tekeste and Marcel Torne and Kyle Vedder and Quan Vuong and Anna Walling and Haohuan Wang and Jason Wang and XuDong Wang and Chris Whalen and Samuel Whitmore and Blake Williams and Charles Xu and Sukwon Yoo and Lili Yu and Wuming Zhang and Zhuoyang Zhang and Ury Zhilinsky},\n year = {2026},\n url = {https://arxiv.org/abs/2604.15483}\n}", "preview": "assets/papers/pi07.png", "previewSourceUrl": "https://arxiv.org/pdf/2604.15483" }, "relatedIds": [ "openpi", "article-pi0", "pi-star06" ], "addedAt": "2026-09-15", "updatedAt": "2026-09-15", "publicationDate": "2026-04-16", "influenceSources": [ "https://www.pi.website/blog/pi07" ], "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7155019052", "updatedAt": "2026-10-12", "recordTitle": "$π_{0.7}$: a Steerable Generalist Robotic Foundation Model with Emergent Capabilities" } } }, { "id": "pi-star06", "institutions": [ { "name": "Physical Intelligence", "url": "https://www.physicalintelligence.company/", "logo": "assets/institutions/physical-intelligence.svg", "mark": "PI" } ], "name": "π*0.6 / RECAP", "title": "π*0.6: a VLA That Learns From Experience", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "arXiv · 2025.11", "summary": "Combines demonstrations, autonomous execution, and human corrections to improve vision-language-action models through advantage-conditioned reinforcement learning.", "tags": [ "Vision-language-action", "Reinforcement learning" ], "environment": "Multitask real-world robot manipulation.", "interface": "VLA action execution with experience and corrective feedback.", "limits": "Requires real-world data collection and human corrections. The reported results do not imply entirely intervention-free self-improvement.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2511.14759", "project": "https://www.pi.website/blog/pistar06", "pdf": "https://arxiv.org/pdf/2511.14759" }, "rank": 10, "publicationDate": "2025-11-18", "selectionTrack": "recent-arxiv", "selectionReason": "Editorial selection: official multi-task results on learning from real-robot experience support a significant direction from demonstration learning toward execution feedback. Publicity and GitHub stars are not treated as research-quality measures.", "influenceSources": [ "https://www.pi.website/blog/pistar06" ], "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2511.14759", "checkedAt": "2026-09-14", "authors": [ "Physical Intelligence", "Ali Amin", "Raichelle Aniceto", "Ashwin Balakrishna", "Kevin Black", "Ken Conley", "Grace Connors", "James Darpinian", "Karan Dhabalia", "Jared DiCarlo", "Danny Driess", "Michael Equi", "Adnan Esmail", "Yunhao Fang", "Chelsea Finn", "Catherine Glossop", "Thomas Godden", "Ivan Goryachev", "Lachy Groom", "Hunter Hancock", "Karol Hausman", "Gashon Hussein", "Brian Ichter", "Szymon Jakubczak", "Rowan Jen", "Tim Jones", "Ben Katz", "Liyiming Ke", "Chandra Kuchi", "Marinda Lamb", "Devin LeBlanc", "Sergey Levine", "Adrian Li-Bell", "Yao Lu", "Vishnu Mano", "Mohith Mothukuri", "Suraj Nair", "Karl Pertsch", "Allen Z. Ren", "Charvi Sharma", "Lucy Xiaoyang Shi", "Laura Smith", "Jost Tobias Springenberg", "Kyle Stachowicz", "Will Stoeckle", "Alex Swerdlow", "James Tanner", "Marcel Torne", "Quan Vuong", "Anna Walling", "Haohuan Wang", "Blake Williams", "Sukwon Yoo", "Lili Yu", "Ury Zhilinsky", "Zhiyuan Zhou" ], "date": "2025-11-18", "dateLabel": "First arXiv submission", "arxivId": "2511.14759", "bibtex": "@misc{pi-star06,\n title = {π*0.6: a VLA That Learns From Experience},\n author = {{Physical Intelligence} and Ali Amin and Raichelle Aniceto and Ashwin Balakrishna and Kevin Black and Ken Conley and Grace Connors and James Darpinian and Karan Dhabalia and Jared DiCarlo and Danny Driess and Michael Equi and Adnan Esmail and Yunhao Fang and Chelsea Finn and Catherine Glossop and Thomas Godden and Ivan Goryachev and Lachy Groom and Hunter Hancock and Karol Hausman and Gashon Hussein and Brian Ichter and Szymon Jakubczak and Rowan Jen and Tim Jones and Ben Katz and Liyiming Ke and Chandra Kuchi and Marinda Lamb and Devin LeBlanc and Sergey Levine and Adrian Li-Bell and Yao Lu and Vishnu Mano and Mohith Mothukuri and Suraj Nair and Karl Pertsch and Allen Z. Ren and Charvi Sharma and Lucy Xiaoyang Shi and Laura Smith and Jost Tobias Springenberg and Kyle Stachowicz and Will Stoeckle and Alex Swerdlow and James Tanner and Marcel Torne and Quan Vuong and Anna Walling and Haohuan Wang and Blake Williams and Sukwon Yoo and Lili Yu and Ury Zhilinsky and Zhiyuan Zhou},\n year = {2025},\n url = {https://arxiv.org/abs/2511.14759}\n}", "preview": "assets/papers/pi-star06.png", "previewSourceUrl": "https://arxiv.org/pdf/2511.14759" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4417215404", "updatedAt": "2026-09-15", "recordTitle": "$π^{*}_{0.6}$: a VLA That Learns From Experience" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-15", "relatedIds": [ "openpi", "article-pi0", "pi0-paper", "pi0-fast", "pi05", "pi05-ki", "pi05-ego", "pi07" ] }, { "id": "lerobot", "name": "LeRobot", "title": "Robot learning: datasets, policies and hardware", "domains": [ "embodied" ], "kinds": [ "projects" ], "year": null, "venue": "Robot learning toolkit", "summary": "Integrates robot data collection, dataset formats, policy training, and hardware interfaces for experiments from demonstration to deployment.", "tags": [ "Robot learning", "Imitation learning" ], "environment": "Real robots and robot-learning datasets.", "interface": "Python robot interfaces, training utilities, and data tools.", "limits": "Install model-specific drivers and calibrate the hardware. Check dataset, model, and code licenses separately.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/huggingface/lerobot" }, "rank": 17, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 28279, "source": "GitHub", "sourceUrl": "https://github.com/huggingface/lerobot/stargazers", "repository": "huggingface/lerobot", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "openpi", "name": "OpenPI", "title": "Physical Intelligence open robot policy tooling", "domains": [ "embodied" ], "kinds": [ "projects" ], "year": null, "venue": "Policy training toolkit", "summary": "Provides training, inference, and adaptation code for Physical Intelligence robot policies, connecting models to task data.", "tags": [ "Vision-language-action", "Robot learning" ], "environment": "Robot policy training and inference.", "interface": "Python, model serving, and dataset configuration.", "limits": "Prepare model weights and hardware dependencies for the chosen release. Do not assume every component of a related paper has been released.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/Physical-Intelligence/openpi" }, "rank": 94, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 14293, "source": "GitHub", "sourceUrl": "https://github.com/Physical-Intelligence/openpi/stargazers", "repository": "Physical-Intelligence/openpi", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-15", "relatedIds": [ "pi0-paper", "pi0-fast", "pi05", "pi05-ki", "pi05-ego", "pi07", "pi-star06" ] }, { "id": "gymnasium", "name": "Gymnasium", "title": "A standard API for reinforcement learning environments", "domains": [ "game", "embodied" ], "kinds": [ "projects" ], "year": null, "venue": "Environment API", "summary": "Provides a common environment interface for single-agent reinforcement learning experiments.", "tags": [ "Reinforcement learning", "Benchmarks" ], "environment": "Diverse single-agent environments.", "interface": "Python reset / step API.", "limits": "A standardized interface does not automatically align task difficulty, observation access, or rewards across environments.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/Farama-Foundation/Gymnasium" }, "rank": 95, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 12698, "source": "GitHub", "sourceUrl": "https://github.com/Farama-Foundation/Gymnasium/stargazers", "repository": "Farama-Foundation/Gymnasium", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "pettingzoo", "name": "PettingZoo", "title": "Multi-agent reinforcement learning APIs", "domains": [ "game" ], "kinds": [ "projects" ], "year": null, "venue": "Multi-agent API", "summary": "Provides multi-agent environment interfaces and reference tasks with turn-based and parallel execution.", "tags": [ "Multi-agent learning", "Reinforcement learning" ], "environment": "Cooperative, competitive, and mixed tasks.", "interface": "AEC and Parallel APIs.", "limits": "Specify communication mechanisms, observability, opponents, and team-reward settings.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/Farama-Foundation/PettingZoo" }, "rank": 96, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 3535, "source": "GitHub", "sourceUrl": "https://github.com/Farama-Foundation/PettingZoo/stargazers", "repository": "Farama-Foundation/PettingZoo", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "genesis", "name": "Genesis", "title": "General-purpose robotics simulation", "domains": [ "embodied" ], "kinds": [ "projects" ], "year": null, "venue": "Physics simulator", "summary": "Provides general-purpose simulation infrastructure for robotics and embodied learning experiments with physical interaction.", "tags": [ "Robot learning", "Environment generation" ], "environment": "Robot simulation scenes.", "interface": "Simulation and control interfaces.", "limits": "Physics, hardware compatibility, and feature support depend on the release. Simulation does not establish validated sim-to-real transfer.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/Genesis-Embodied-AI/genesis-world" }, "rank": 97, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 30084, "source": "GitHub", "sourceUrl": "https://github.com/Genesis-Embodied-AI/genesis-world/stargazers", "repository": "Genesis-Embodied-AI/genesis-world", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "vizdoom", "name": "ViZDoom", "title": "Doom-based visual reinforcement learning environments", "domains": [ "game" ], "kinds": [ "benchmarks" ], "year": null, "venue": "Game research environment", "summary": "Provides configurable first-person visual tasks based on DOOM for navigation, survival, and shooting research.", "tags": [ "Benchmarks", "Reinforcement learning" ], "environment": "DOOM.", "interface": "Visual observations, game variables, and discrete actions.", "limits": "Specify access to game variables, frame skipping, and the action set. The underlying environment is not subject to the paper publication window.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/Farama-Foundation/ViZDoom" }, "rank": 98, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 2100, "source": "GitHub", "sourceUrl": "https://github.com/Farama-Foundation/ViZDoom/stargazers", "repository": "Farama-Foundation/ViZDoom", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "habitat", "name": "Habitat-Lab", "title": "Embodied AI tasks and environments", "domains": [ "embodied" ], "kinds": [ "benchmarks" ], "year": null, "venue": "Embodied research environment", "summary": "Provides embodied tasks, environment configuration, and training/evaluation components for navigation and interaction research.", "tags": [ "Benchmarks", "Planning & memory" ], "environment": "3D indoor environments.", "interface": "Sensor observations and task-action interfaces.", "limits": "Scene datasets and the simulator must be configured together. Evaluation protocols depend on the task.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/facebookresearch/habitat-lab" }, "rank": 99, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 3162, "source": "GitHub", "sourceUrl": "https://github.com/facebookresearch/habitat-lab/stargazers", "repository": "facebookresearch/habitat-lab", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "ai2thor", "name": "AI2-THOR", "title": "Interactive 3D environments for visual AI", "domains": [ "embodied" ], "kinds": [ "benchmarks" ], "year": null, "venue": "Interactive environment", "summary": "Provides interactive indoor scenes and objects for visual navigation, object interaction, and embodied planning.", "tags": [ "Benchmarks", "Planning & memory" ], "environment": "Interactive 3D indoor scenes.", "interface": "Visual observations and environment action APIs.", "limits": "Discrete object interaction differs from real joint-level control. State the exact task definition.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/allenai/ai2thor" }, "rank": 100, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 1812, "source": "GitHub", "sourceUrl": "https://github.com/allenai/ai2thor/stargazers", "repository": "allenai/ai2thor", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "maniskill", "name": "ManiSkill", "title": "GPU-parallelized robotics simulation and benchmarks", "domains": [ "embodied" ], "kinds": [ "benchmarks" ], "year": null, "venue": "Manipulation benchmark", "summary": "Provides parallel simulation, task environments, and evaluation tools for robot manipulation skills.", "tags": [ "Benchmarks", "Robot learning" ], "environment": "Simulated robot manipulation.", "interface": "Simulated observations, actions, and skill tasks.", "limits": "Specify the environment version, robot, control mode, and training data. Simulation throughput is not real-world control performance.", "status": "selected", "evidence": "Official repository purpose and interfaces checked; not installed locally.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "code": "https://github.com/mani-skill/ManiSkill" }, "rank": 101, "selectionTrack": "infrastructure", "selectionReason": "Standalone open-source tool or environment supporting research and reproduction; excluded from paper-window statistics.", "metrics": { "github": { "value": 3429, "source": "GitHub", "sourceUrl": "https://github.com/mani-skill/ManiSkill/stargazers", "repository": "mani-skill/ManiSkill", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-s08", "name": "X-Ego", "title": "X-Ego: Acquiring Team-Level Tactical Situational Awareness via Cross-Egocentric Contrastive Video Representation Learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "Candidate · arXiv", "summary": "Studies whether synchronized teammate first-person views during training help a single-view model infer otherwise hidden team situations.", "tags": [ "Pending review" ], "environment": "Counter-Strike 2, de_mirage.", "interface": "Training uses synchronized first-person video from multiple players. Inference can use one view to classify current player regions; it does not output combat actions.", "limits": "Separate training and test observation access. Location classification is not win rate; some encoders degrade, and round-level splits do not establish cross-player or cross-map generalization.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2510.19150", "pdf": "https://arxiv.org/pdf/2510.19150" }, "rank": 72, "sourceId": "S08", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2510.19150", "checkedAt": "2026-09-14", "authors": [ "Yunzhe Wang", "Soham Hans", "Volkan Ustun" ], "date": "2025-10-22", "dateLabel": "First arXiv submission", "arxivId": "2510.19150", "bibtex": "@misc{candidate-s08,\n title = {X-Ego: Acquiring Team-Level Tactical Situational Awareness via Cross-Egocentric Contrastive Video Representation Learning},\n author = {Yunzhe Wang and Soham Hans and Volkan Ustun},\n year = {2025},\n url = {https://arxiv.org/abs/2510.19150}\n}", "preview": "assets/papers/candidate-s08.png", "previewSourceUrl": "https://arxiv.org/pdf/2510.19150" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4416246481", "updatedAt": "2026-10-19", "recordTitle": "X-Ego: Acquiring Team-Level Tactical Situational Awareness via Cross-Egocentric Contrastive Video Representation Learning" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-p01", "name": "Game-TARS", "title": "Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2025, "venue": "Candidate · arXiv", "summary": "Studies generalist cross-game control with a shared keyboard/mouse representation, addressing latency and repetitive actions in stepwise multimodal reasoning.", "tags": [ "Pending review" ], "environment": "ViZDoom, Minecraft, Web 3D, and MiniWorld.", "interface": "Multimodal images and trajectories map to native keyboard/mouse action sequences. Exact FPS observation access and execution timing need verification.", "limits": "Paused browser execution and best-of-five results do not establish real-time FPS performance. Keep Doom returns, Minecraft success rates, and online combat ablations separate, with data and scale differences noted.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2510.23691", "pdf": "https://arxiv.org/pdf/2510.23691" }, "rank": 73, "sourceId": "P01", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2510.23691", "checkedAt": "2026-09-14", "authors": [ "Zihao Wang", "Xujing Li", "Yining Ye", "Junjie Fang", "Haoming Wang", "Longxiang Liu", "Shihao Liang", "Junting Lu", "Zhiyong Wu", "Jiazhan Feng", "Wanjun Zhong", "Zili Li", "Yu Wang", "Yu Miao", "Bo Zhou", "Yuanfan Li", "Hao Wang", "Zhongkai Zhao", "Faming Wu", "Zhengxuan Jiang", "Weihao Tan", "Heyuan Yao", "Shi Yan", "Xiangyang Li", "Yitao Liang", "Yujia Qin", "Guang Shi" ], "date": "2025-10-27", "dateLabel": "First arXiv submission", "arxivId": "2510.23691", "bibtex": "@misc{candidate-p01,\n title = {Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents},\n author = {Zihao Wang and Xujing Li and Yining Ye and Junjie Fang and Haoming Wang and Longxiang Liu and Shihao Liang and Junting Lu and Zhiyong Wu and Jiazhan Feng and Wanjun Zhong and Zili Li and Yu Wang and Yu Miao and Bo Zhou and Yuanfan Li and Hao Wang and Zhongkai Zhao and Faming Wu and Zhengxuan Jiang and Weihao Tan and Heyuan Yao and Shi Yan and Xiangyang Li and Yitao Liang and Yujia Qin and Guang Shi},\n year = {2025},\n url = {https://arxiv.org/abs/2510.23691}\n}", "preview": "assets/papers/candidate-p01.png", "previewSourceUrl": "https://arxiv.org/pdf/2510.23691" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7102387721", "updatedAt": "2026-10-19", "recordTitle": "Game-TARS: Pretrained Foundation Models for Scalable Generalist Multimodal Game Agents" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-p02", "name": "GameWAM", "title": "GameWAM: A World Action Model for Video Games", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Combines future-frame supervision with action learning to study whether modeling world changes improves game control, and why generation noise causes camera drift.", "tags": [ "Pending review" ], "environment": "ViZDoom and Minecraft MCU.", "interface": "ViZDoom inputs include images and position, orientation, and health state. In-map APPO expert trajectories train keyboard/mouse action chunks with receding-horizon execution.", "limits": "This is neither vision-only nor zero-shot FPS control. Data, inputs, and architectures differ from baselines, allowing only system-level comparisons. Minecraft ablations are not Doom results; execution steps are not elapsed time.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2608.26200", "pdf": "https://arxiv.org/pdf/2608.26200" }, "rank": 74, "sourceId": "P02", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2608.26200", "checkedAt": "2026-09-14", "authors": [ "Yuncheng Guo", "Zhanqiu Zhang", "Yiwen Guo", "Weijia Li" ], "date": "2026-08-25", "dateLabel": "First arXiv submission", "arxivId": "2608.26200", "bibtex": "@misc{candidate-p02,\n title = {GameWAM: A World Action Model for Video Games},\n author = {Yuncheng Guo and Zhanqiu Zhang and Yiwen Guo and Weijia Li},\n year = {2026},\n url = {https://arxiv.org/abs/2608.26200}\n}", "preview": "assets/papers/candidate-p02.png", "previewSourceUrl": "https://arxiv.org/pdf/2608.26200" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7204581771", "updatedAt": "2026-10-19", "recordTitle": "GameWAM: A World Action Model for Video Games" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-t03", "name": "See, Symbolize, Act", "title": "See, Symbolize, Act: Grounding VLMs with Spatial Representations for Better Gameplay", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Studies whether visual-language game-agent failures arise from perception or decision-making, and when symbolic spatial information helps.", "tags": [ "Pending review" ], "environment": "ViZDoom defend_the_center; also Atari and AI2-THOR.", "interface": "Compares images, self-extracted object symbols, and engine labels-buffer ground truth; outputs rotation and ATTACK in defend_the_center.", "limits": "Engine ground truth provides extra information, so gains cannot be attributed entirely to reasoning. The fixed-position single-scenario setting has limited variance evidence. An AAAI Bridge is not the main conference.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2603.11601", "pdf": "https://arxiv.org/pdf/2603.11601" }, "rank": 75, "sourceId": "T03", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2603.11601", "checkedAt": "2026-09-14", "authors": [ "Ashish Baghel", "Paras Chopra" ], "date": "2026-03-12", "dateLabel": "First arXiv submission", "arxivId": "2603.11601", "bibtex": "@misc{candidate-t03,\n title = {See, Symbolize, Act: Grounding VLMs with Spatial Representations for Better Gameplay},\n author = {Ashish Baghel and Paras Chopra},\n year = {2026},\n url = {https://arxiv.org/abs/2603.11601}\n}", "preview": "assets/papers/candidate-t03.png", "previewSourceUrl": "https://arxiv.org/pdf/2603.11601" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7135428496", "updatedAt": "2026-09-21", "recordTitle": "See, Symbolize, Act: Grounding VLMs with Spatial Representations for Better Gameplay" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-t04", "name": "Playing DOOM with 1.3M Parameters", "title": "Playing DOOM with 1.3M Parameters: Specialized Small Models vs Large Language Models for Real-Time Game Control", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Tests task-specific small models for low-latency control in real-time shooting, comparing them with general language models.", "tags": [ "Pending review" ], "environment": "ViZDoom defend_the_center.", "interface": "ASCII images and quantized engine depth feed supervised learning from human demonstrations for compound turning/firing actions in a fixed defense scenario.", "limits": "Task-specific supervised learning differs from zero-shot LLM use. Align episode budgets, frame skipping, hardware, and observation access. Random neighboring-frame splits do not demonstrate cross-scene generalization.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2604.07385", "pdf": "https://arxiv.org/pdf/2604.07385" }, "rank": 76, "sourceId": "T04", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2604.07385", "checkedAt": "2026-09-14", "authors": [ "David Golchinfar", "Daryoush Vaziri", "Alexander Marquardt" ], "date": "2026-04-08", "dateLabel": "First arXiv submission", "arxivId": "2604.07385", "bibtex": "@misc{candidate-t04,\n title = {Playing DOOM with 1.3M Parameters: Specialized Small Models vs Large Language Models for Real-Time Game Control},\n author = {David Golchinfar and Daryoush Vaziri and Alexander Marquardt},\n year = {2026},\n url = {https://arxiv.org/abs/2604.07385}\n}", "preview": "assets/papers/candidate-t04.png", "previewSourceUrl": "https://arxiv.org/pdf/2604.07385" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7153669982", "updatedAt": "2026-09-21", "recordTitle": "Playing DOOM with 1.3M Parameters: Specialized Small Models vs Large Language Models for Real-Time Game Control" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-t05", "name": "Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence", "title": "Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Studies how an XR assistant can identify targets a player intends to engage but cannot shoot in time, beyond learning only from completed shots.", "tags": [ "Pending review" ], "environment": "Unity XR SpaceShooter: distinguish friendly/enemy ships and protect a mothership.", "interface": "Real-user shooting behavior, fixation-locked EEG, and visual information produce target-direction cues. The human performs the actual shooting.", "limits": "Cue convergence and belief convergence are different. Small exploratory samples do not establish all between-group differences; results are not autonomous-control win rates. The author's claimed acceptance still requires verification.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2608.30369", "pdf": "https://arxiv.org/pdf/2608.30369" }, "rank": 77, "sourceId": "T05", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2608.30369", "checkedAt": "2026-09-14", "authors": [ "Ziheng Li", "Xichen He", "Haoyan Chen", "Charlie Zou", "Sheng Bai", "Benjamin Yang", "Mengyuan Wu", "Jake Ledner", "Yi-Jie Cheng", "Akito Yamauchi", "Dishita G Turakhia", "Steven Feiner", "Paul Sajda" ], "date": "2026-08-31", "dateLabel": "First arXiv submission", "arxivId": "2608.30369", "bibtex": "@misc{candidate-t05,\n title = {Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence},\n author = {Ziheng Li and Xichen He and Haoyan Chen and Charlie Zou and Sheng Bai and Benjamin Yang and Mengyuan Wu and Jake Ledner and Yi-Jie Cheng and Akito Yamauchi and Dishita G Turakhia and Steven Feiner and Paul Sajda},\n year = {2026},\n url = {https://arxiv.org/abs/2608.30369}\n}", "preview": "assets/papers/candidate-t05.png", "previewSourceUrl": "https://arxiv.org/pdf/2608.30369" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7204885827", "updatedAt": "2026-10-19", "recordTitle": "Augmenting Human Performance with an XR Agent Learning from Online Behavior and BCI Evidence" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-t06", "name": "Procedural Generation of First Person Shooter Maps using Map-Elites", "title": "Procedural Generation of First Person Shooter Maps using Map-Elites", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Uses quality-diversity search to generate FPS maps with different layouts and combat rhythms, rather than a single balanced map.", "tags": [ "Pending review" ], "environment": "Unity Project Arena, Duel 1v1 deathmatch.", "interface": "Map representations and search parameters produce layouts, evaluated by predefined bots in 1v1 deathmatch.", "limits": "Procedural generation here is not a neural world model. Kill entropy measures balance, not enjoyment; bot skill, weapons, and search budgets constrain conclusions about human experience.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2605.30570", "pdf": "https://arxiv.org/pdf/2605.30570" }, "rank": 78, "sourceId": "T06", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2605.30570", "checkedAt": "2026-09-14", "authors": [ "Simone de Donato", "Pier Luca Lanzi", "Daniele Loiacono" ], "date": "2026-05-28", "dateLabel": "First arXiv submission", "arxivId": "2605.30570", "bibtex": "@misc{candidate-t06,\n title = {Procedural Generation of First Person Shooter Maps using Map-Elites},\n author = {Simone de Donato and Pier Luca Lanzi and Daniele Loiacono},\n year = {2026},\n url = {https://arxiv.org/abs/2605.30570}\n}", "preview": "assets/papers/candidate-t06.png", "previewSourceUrl": "https://arxiv.org/pdf/2605.30570" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7163287888", "updatedAt": "2026-09-21", "recordTitle": "Procedural Generation of First Person Shooter Maps using Map-Elites" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "candidate-t07", "name": "Account Consistency from Gameplay Traces", "title": "Account Consistency from Gameplay Traces: Same-Player Verification in Counter-Strike 2", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2026, "venue": "Candidate · arXiv", "summary": "Uses gameplay behavior fingerprints to assess whether a CS2 account's current match was played by its historical operator.", "tags": [ "Pending review" ], "environment": "Real Counter-Strike 2 match demos.", "interface": "Crosshair, movement, firing, economy, and timing traces produce a cross-match same-operator score. The system does not control the game.", "limits": "AUC is not deployment accuracy. Fix query sets and history size before comparing; same-person verification is not cheating detection. Contaminated history and noisy identity records affect conclusions.", "status": "candidate", "evidence": "Knowledge-base lead; verification for curated inclusion is incomplete.", "sourceType": "knowledge", "checkedAt": "2026-09-14", "links": { "paper": "https://arxiv.org/abs/2608.24893", "pdf": "https://arxiv.org/pdf/2608.24893" }, "rank": 79, "sourceId": "T07", "selectionTrack": "candidate", "selectionReason": "Recent preprint lead; official acceptance or evidence of significance needs further verification.", "paperType": "method", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2608.24893", "checkedAt": "2026-09-14", "authors": [ "Xuchen Zhang" ], "date": "2026-06-20", "dateLabel": "First arXiv submission", "arxivId": "2608.24893", "bibtex": "@misc{candidate-t07,\n title = {Account Consistency from Gameplay Traces: Same-Player Verification in Counter-Strike 2},\n author = {Xuchen Zhang},\n year = {2026},\n url = {https://arxiv.org/abs/2608.24893}\n}", "preview": "assets/papers/candidate-t07.png", "previewSourceUrl": "https://arxiv.org/pdf/2608.24893" }, "metrics": { "citations": { "value": 0, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W7204671282", "updatedAt": "2026-09-21", "recordTitle": "Account Consistency from Gameplay Traces: Same-Player Verification in Counter-Strike 2" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "survey-foundation-robotics", "name": "Foundation models in robotics", "title": "Foundation models in robotics: Applications, challenges, and the future", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2025, "venue": "IJRR 2025", "summary": "Reviews how foundation models enter robot perception, planning, and control, with a research agenda spanning data, uncertainty, safety, latency, and reproducibility.", "tags": [ "Surveys", "Robot learning" ], "environment": "Robot manipulation, navigation, and related embodied tasks.", "interface": "Survey of model roles and robot system interfaces.", "limits": "A survey organizes research evidence; it is not a deployable policy or an independent benchmark.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://journals.sagepub.com/doi/10.1177/02783649241281508", "pdf": "https://journals.sagepub.com/history/e3eb5469-f176-4c99-8231-efe683fe6399/02783649241281508.17904978.pdf" }, "rank": 160, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "survey", "credit": "Roya Firoozi and collaborators", "bibliography": { "source": "Crossref", "sourceUrl": "https://doi.org/10.1177/02783649241281508", "checkedAt": "2026-09-14", "authors": [ "Roya Firoozi", "Johnathan Tucker", "Stephen Tian", "Anirudha Majumdar", "Jiankai Sun", "Weiyu Liu", "Yuke Zhu", "Shuran Song", "Ashish Kapoor", "Karol Hausman", "Brian Ichter", "Danny Driess", "Jiajun Wu", "Cewu Lu", "Mac Schwager" ], "date": "2024-09-25", "dateLabel": "Publication date (Crossref)", "doi": "10.1177/02783649241281508", "bibtex": "@misc{survey-foundation-robotics,\n title = {Foundation models in robotics: Applications, challenges, and the future},\n author = {Roya Firoozi and Johnathan Tucker and Stephen Tian and Anirudha Majumdar and Jiankai Sun and Weiyu Liu and Yuke Zhu and Shuran Song and Ashish Kapoor and Karol Hausman and Brian Ichter and Danny Driess and Jiajun Wu and Cewu Lu and Mac Schwager},\n year = {2024},\n url = {https://journals.sagepub.com/doi/10.1177/02783649241281508},\n doi = {10.1177/02783649241281508}\n}", "preview": "assets/papers/survey-foundation-robotics.png", "previewSourceUrl": "https://journals.sagepub.com/history/e3eb5469-f176-4c99-8231-efe683fe6399/02783649241281508.17904978.pdf" }, "metrics": { "citations": { "value": 218, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402890475", "updatedAt": "2026-10-12", "recordTitle": "Foundation models in robotics: Applications, challenges, and the future" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "survey-foundation-manipulation", "institutions": [ { "name": "Samsung R&D Institute China-Beijing", "url": "https://research.samsung.com/srcc", "logo": "assets/institutions/samsung.svg", "mark": "SR" }, { "name": "Beijing University of Posts and Telecommunications", "url": "https://www.bupt.edu.cn/", "mark": "BUPT" }, { "name": "Tsinghua University", "url": "https://www.tsinghua.edu.cn/", "logo": "assets/institutions/tsinghua.png", "mark": "THU" }, { "name": "Southeast University", "url": "https://www.seu.edu.cn/", "mark": "SEU" }, { "name": "Universität Hamburg", "url": "https://www.uni-hamburg.de/", "logo": "assets/institutions/hamburg.png", "mark": "UHH" } ], "name": "Foundation models for manipulation", "title": "What foundation models can bring for robot learning in manipulation: A survey", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2026, "venue": "IJRR 2026", "summary": "Organizes foundation models into a modular framework for general manipulation and examines the opportunities and limits of each role.", "tags": [ "Surveys", "Robot learning" ], "environment": "Manipulation in unstructured environments.", "interface": "Survey of perception, learning, planning, and manipulation modules.", "limits": "The framework is a synthesis of research directions, not proof that one foundation model solves general manipulation.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://journals.sagepub.com/doi/10.1177/02783649251390579", "pdf": "https://arxiv.org/pdf/2404.18201" }, "rank": 161, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "survey", "credit": "Dingzhe Li and collaborators", "publicationNote": "First published online on 2025-11-20; the journal issue is dated 2026.", "bibliography": { "source": "Crossref", "sourceUrl": "https://doi.org/10.1177/02783649251390579", "checkedAt": "2026-09-14", "authors": [ "Dingzhe Li", "Yixiang Jin", "YuHao Sun", "Yong A", "Hongze Yu", "Jun Shi", "Xiaoshuai Hao", "Peng Hao", "Huaping Liu", "Xiang Li", "Xinde Li", "Fuchun Sun", "Jianwei Zhang", "Bin Fang" ], "date": "2025-11-20", "dateLabel": "Publication date (Crossref)", "doi": "10.1177/02783649251390579", "bibtex": "@misc{survey-foundation-manipulation,\n title = {What foundation models can bring for robot learning in manipulation: A survey},\n author = {Dingzhe Li and Yixiang Jin and YuHao Sun and Yong A and Hongze Yu and Jun Shi and Xiaoshuai Hao and Peng Hao and Huaping Liu and Xiang Li and Xinde Li and Fuchun Sun and Jianwei Zhang and Bin Fang},\n year = {2025},\n url = {https://journals.sagepub.com/doi/10.1177/02783649251390579},\n doi = {10.1177/02783649251390579}\n}", "preview": "assets/papers/survey-foundation-manipulation.png", "previewVersion": "arXiv version", "previewSourceUrl": "https://arxiv.org/pdf/2404.18201" }, "metrics": { "citations": { "value": 6, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4416422407", "updatedAt": "2026-09-28", "recordTitle": "What foundation models can bring for robot learning in manipulation: A survey" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "minedojo", "institutions": [ { "name": "NVIDIA", "url": "https://www.nvidia.com/", "logo": "assets/institutions/nvidia.png", "mark": "NV" } ], "name": "MineDojo", "title": "MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge", "domains": [ "game" ], "kinds": [ "datasets", "papers" ], "year": 2022, "venue": "NeurIPS 2022 D&B", "summary": "Combines Minecraft tasks with a multimodal knowledge collection of videos, tutorials, wiki pages, and discussions for open-ended agent learning.", "tags": [ "Datasets", "Language & action" ], "environment": "Minecraft open-ended tasks and internet knowledge.", "interface": "Simulation API, video-language rewards, and multimodal knowledge access.", "limits": "The knowledge collection and simulator are distinct resources. Check access terms and task definitions; broad task coverage does not imply a general competent player.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/74a67268c5cc5910f64938cac4526a90-Abstract.html", "code": "https://github.com/MineDojo/MineDojo", "data": "https://minedojo.org/", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2022/file/74a67268c5cc5910f64938cac4526a90-Paper-Datasets_and_Benchmarks.pdf" }, "rank": 162, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "dataset", "credit": "MineDojo research team", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/74a67268c5cc5910f64938cac4526a90-Abstract.html", "checkedAt": "2026-09-14", "authors": [ "Linxi Fan", "Guanzhi Wang", "Yunfan Jiang", "Ajay Mandlekar", "Yuncong Yang", "Haoyi Zhu", "Andrew Tang", "De-An Huang", "Yuke Zhu", "Anima Anandkumar" ], "date": "2022-12-06", "dateLabel": "Publisher record date", "doi": "10.52202/068431-1333", "bibtex": "@misc{minedojo,\n title = {MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge},\n author = {Linxi Fan and Guanzhi Wang and Yunfan Jiang and Ajay Mandlekar and Yuncong Yang and Haoyi Zhu and Andrew Tang and De-An Huang and Yuke Zhu and Anima Anandkumar},\n year = {2022},\n url = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/74a67268c5cc5910f64938cac4526a90-Abstract.html},\n doi = {10.52202/068431-1333}\n}", "preview": "assets/papers/minedojo.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/file/74a67268c5cc5910f64938cac4526a90-Paper-Datasets_and_Benchmarks.pdf" }, "metrics": { "citations": { "value": 59, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4283218993", "updatedAt": "2026-09-21", "recordTitle": "MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge" }, "github": { "value": 2276, "source": "GitHub", "sourceUrl": "https://github.com/MineDojo/MineDojo/stargazers", "repository": "MineDojo/MineDojo", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "open-x-embodiment", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "Open X-Embodiment", "title": "Open X-Embodiment: Robotic Learning Datasets and RT-X Models", "domains": [ "embodied" ], "kinds": [ "datasets", "papers" ], "year": 2024, "venue": "ICRA 2024", "summary": "Brings together diverse robot trajectories in shared formats to study cross-robot transfer and generalist policies.", "tags": [ "Datasets", "Vision-language-action" ], "environment": "Multiple real robot embodiments and manipulation settings.", "interface": "Standardized trajectory data and RT-X action models.", "limits": "Datasets retain their individual collection conditions and licenses. Shared formats do not remove embodiment or evaluation differences.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://doi.org/10.1109/ICRA57147.2024.10611477", "code": "https://github.com/google-deepmind/open_x_embodiment", "data": "https://robotics-transformer-x.github.io/", "pdf": "https://arxiv.org/pdf/2310.08864" }, "rank": 163, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "dataset", "credit": "Open X-Embodiment Collaboration", "bibliography": { "source": "arXiv", "sourceUrl": "https://arxiv.org/abs/2310.08864", "checkedAt": "2026-09-14", "authors": [ "Embodiment Collaboration", "Abby O'Neill", "Abdul Rehman", "Abhinav Gupta", "Abhiram Maddukuri", "Abhishek Gupta", "Abhishek Padalkar", "Abraham Lee", "Acorn Pooley", "Agrim Gupta", "Ajay Mandlekar", "Ajinkya Jain", "Albert Tung", "Alex Bewley", "Alex Herzog", "Alex Irpan", "Alexander Khazatsky", "Anant Rai", "Anchit Gupta", "Andrew Wang", "Andrey Kolobov", "Anikait Singh", "Animesh Garg", "Aniruddha Kembhavi", "Annie Xie", "Anthony Brohan", "Antonin Raffin", "Archit Sharma", "Arefeh Yavary", "Arhan Jain", "Ashwin Balakrishna", "Ayzaan Wahid", "Ben Burgess-Limerick", "Beomjoon Kim", "Bernhard Schölkopf", "Blake Wulfe", "Brian Ichter", "Cewu Lu", "Charles Xu", "Charlotte Le", "Chelsea Finn", "Chen Wang", "Chenfeng Xu", "Cheng Chi", "Chenguang Huang", "Christine Chan", "Christopher Agia", "Chuer Pan", "Chuyuan Fu", "Coline Devin", "Danfei Xu", "Daniel Morton", "Danny Driess", "Daphne Chen", "Deepak Pathak", "Dhruv Shah", "Dieter Büchler", "Dinesh Jayaraman", "Dmitry Kalashnikov", "Dorsa Sadigh", "Edward Johns", "Ethan Foster", "Fangchen Liu", "Federico Ceola", "Fei Xia", "Feiyu Zhao", "Felipe Vieira Frujeri", "Freek Stulp", "Gaoyue Zhou", "Gaurav S. Sukhatme", "Gautam Salhotra", "Ge Yan", "Gilbert Feng", "Giulio Schiavi", "Glen Berseth", "Gregory Kahn", "Guangwen Yang", "Guanzhi Wang", "Hao Su", "Hao-Shu Fang", "Haochen Shi", "Henghui Bao", "Heni Ben Amor", "Henrik I Christensen", "Hiroki Furuta", "Homanga Bharadhwaj", "Homer Walke", "Hongjie Fang", "Huy Ha", "Igor Mordatch", "Ilija Radosavovic", "Isabel Leal", "Jacky Liang", "Jad Abou-Chakra", "Jaehyung Kim", "Jaimyn Drake", "Jan Peters", "Jan Schneider", "Jasmine Hsu", "Jay Vakil", "Jeannette Bohg", "Jeffrey Bingham", "Jeffrey Wu", "Jensen Gao", "Jiaheng Hu", "Jiajun Wu", "Jialin Wu", "Jiankai Sun", "Jianlan Luo", "Jiayuan Gu", "Jie Tan", "Jihoon Oh", "Jimmy Wu", "Jingpei Lu", "Jingyun Yang", "Jitendra Malik", "João Silvério", "Joey Hejna", "Jonathan Booher", "Jonathan Tompson", "Jonathan Yang", "Jordi Salvador", "Joseph J. Lim", "Junhyek Han", "Kaiyuan Wang", "Kanishka Rao", "Karl Pertsch", "Karol Hausman", "Keegan Go", "Keerthana Gopalakrishnan", "Ken Goldberg", "Kendra Byrne", "Kenneth Oslund", "Kento Kawaharazuka", "Kevin Black", "Kevin Lin", "Kevin Zhang", "Kiana Ehsani", "Kiran Lekkala", "Kirsty Ellis", "Krishan Rana", "Krishnan Srinivasan", "Kuan Fang", "Kunal Pratap Singh", "Kuo-Hao Zeng", "Kyle Hatch", "Kyle Hsu", "Laurent Itti", "Lawrence Yunliang Chen", "Lerrel Pinto", "Li Fei-Fei", "Liam Tan", "Linxi \"Jim\" Fan", "Lionel Ott", "Lisa Lee", "Luca Weihs", "Magnum Chen", "Marion Lepert", "Marius Memmel", "Masayoshi Tomizuka", "Masha Itkina", "Mateo Guaman Castro", "Max Spero", "Maximilian Du", "Michael Ahn", "Michael C. Yip", "Mingtong Zhang", "Mingyu Ding", "Minho Heo", "Mohan Kumar Srirama", "Mohit Sharma", "Moo Jin Kim", "Muhammad Zubair Irshad", "Naoaki Kanazawa", "Nicklas Hansen", "Nicolas Heess", "Nikhil J Joshi", "Niko Suenderhauf", "Ning Liu", "Norman Di Palo", "Nur Muhammad Mahi Shafiullah", "Oier Mees", "Oliver Kroemer", "Osbert Bastani", "Pannag R Sanketi", "Patrick \"Tree\" Miller", "Patrick Yin", "Paul Wohlhart", "Peng Xu", "Peter David Fagan", "Peter Mitrano", "Pierre Sermanet", "Pieter Abbeel", "Priya Sundaresan", "Qiuyu Chen", "Quan Vuong", "Rafael Rafailov", "Ran Tian", "Ria Doshi", "Roberto Martín-Martín", "Rohan Baijal", "Rosario Scalise", "Rose Hendrix", "Roy Lin", "Runjia Qian", "Ruohan Zhang", "Russell Mendonca", "Rutav Shah", "Ryan Hoque", "Ryan Julian", "Samuel Bustamante", "Sean Kirmani", "Sergey Levine", "Shan Lin", "Sherry Moore", "Shikhar Bahl", "Shivin Dass", "Shubham Sonawani", "Shubham Tulsiani", "Shuran Song", "Sichun Xu", "Siddhant Haldar", "Siddharth Karamcheti", "Simeon Adebola", "Simon Guist", "Soroush Nasiriany", "Stefan Schaal", "Stefan Welker", "Stephen Tian", "Subramanian Ramamoorthy", "Sudeep Dasari", "Suneel Belkhale", "Sungjae Park", "Suraj Nair", "Suvir Mirchandani", "Takayuki Osa", "Tanmay Gupta", "Tatsuya Harada", "Tatsuya Matsushima", "Ted Xiao", "Thomas Kollar", "Tianhe Yu", "Tianli Ding", "Todor Davchev", "Tony Z. Zhao", "Travis Armstrong", "Trevor Darrell", "Trinity Chung", "Vidhi Jain", "Vikash Kumar", "Vincent Vanhoucke", "Vitor Guizilini", "Wei Zhan", "Wenxuan Zhou", "Wolfram Burgard", "Xi Chen", "Xiangyu Chen", "Xiaolong Wang", "Xinghao Zhu", "Xinyang Geng", "Xiyuan Liu", "Xu Liangwei", "Xuanlin Li", "Yansong Pang", "Yao Lu", "Yecheng Jason Ma", "Yejin Kim", "Yevgen Chebotar", "Yifan Zhou", "Yifeng Zhu", "Yilin Wu", "Ying Xu", "Yixuan Wang", "Yonatan Bisk", "Yongqiang Dou", "Yoonyoung Cho", "Youngwoon Lee", "Yuchen Cui", "Yue Cao", "Yueh-Hua Wu", "Yujin Tang", "Yuke Zhu", "Yunchu Zhang", "Yunfan Jiang", "Yunshuang Li", "Yunzhu Li", "Yusuke Iwasawa", "Yutaka Matsuo", "Zehan Ma", "Zhuo Xu", "Zichen Jeff Cui", "Zichen Zhang", "Zipeng Fu", "Zipeng Lin" ], "date": "2023-10-13", "dateLabel": "First arXiv submission", "doi": "10.1109/icra57147.2024.10611477", "arxivId": "2310.08864", "bibtex": "@misc{open-x-embodiment,\n title = {Open X-Embodiment: Robotic Learning Datasets and RT-X Models},\n author = {Embodiment Collaboration and Abby O'Neill and Abdul Rehman and Abhinav Gupta and Abhiram Maddukuri and Abhishek Gupta and Abhishek Padalkar and Abraham Lee and Acorn Pooley and Agrim Gupta and Ajay Mandlekar and Ajinkya Jain and Albert Tung and Alex Bewley and Alex Herzog and Alex Irpan and Alexander Khazatsky and Anant Rai and Anchit Gupta and Andrew Wang and Andrey Kolobov and Anikait Singh and Animesh Garg and Aniruddha Kembhavi and Annie Xie and Anthony Brohan and Antonin Raffin and Archit Sharma and Arefeh Yavary and Arhan Jain and Ashwin Balakrishna and Ayzaan Wahid and Ben Burgess-Limerick and Beomjoon Kim and Bernhard Schölkopf and Blake Wulfe and Brian Ichter and Cewu Lu and Charles Xu and Charlotte Le and Chelsea Finn and Chen Wang and Chenfeng Xu and Cheng Chi and Chenguang Huang and Christine Chan and Christopher Agia and Chuer Pan and Chuyuan Fu and Coline Devin and Danfei Xu and Daniel Morton and Danny Driess and Daphne Chen and Deepak Pathak and Dhruv Shah and Dieter Büchler and Dinesh Jayaraman and Dmitry Kalashnikov and Dorsa Sadigh and Edward Johns and Ethan Foster and Fangchen Liu and Federico Ceola and Fei Xia and Feiyu Zhao and Felipe Vieira Frujeri and Freek Stulp and Gaoyue Zhou and Gaurav S. Sukhatme and Gautam Salhotra and Ge Yan and Gilbert Feng and Giulio Schiavi and Glen Berseth and Gregory Kahn and Guangwen Yang and Guanzhi Wang and Hao Su and Hao-Shu Fang and Haochen Shi and Henghui Bao and Heni Ben Amor and Henrik I Christensen and Hiroki Furuta and Homanga Bharadhwaj and Homer Walke and Hongjie Fang and Huy Ha and Igor Mordatch and Ilija Radosavovic and Isabel Leal and Jacky Liang and Jad Abou-Chakra and Jaehyung Kim and Jaimyn Drake and Jan Peters and Jan Schneider and Jasmine Hsu and Jay Vakil and Jeannette Bohg and Jeffrey Bingham and Jeffrey Wu and Jensen Gao and Jiaheng Hu and Jiajun Wu and Jialin Wu and Jiankai Sun and Jianlan Luo and Jiayuan Gu and Jie Tan and Jihoon Oh and Jimmy Wu and Jingpei Lu and Jingyun Yang and Jitendra Malik and João Silvério and Joey Hejna and Jonathan Booher and Jonathan Tompson and Jonathan Yang and Jordi Salvador and Joseph J. Lim and Junhyek Han and Kaiyuan Wang and Kanishka Rao and Karl Pertsch and Karol Hausman and Keegan Go and Keerthana Gopalakrishnan and Ken Goldberg and Kendra Byrne and Kenneth Oslund and Kento Kawaharazuka and Kevin Black and Kevin Lin and Kevin Zhang and Kiana Ehsani and Kiran Lekkala and Kirsty Ellis and Krishan Rana and Krishnan Srinivasan and Kuan Fang and Kunal Pratap Singh and Kuo-Hao Zeng and Kyle Hatch and Kyle Hsu and Laurent Itti and Lawrence Yunliang Chen and Lerrel Pinto and Li Fei-Fei and Liam Tan and Linxi \"Jim\" Fan and Lionel Ott and Lisa Lee and Luca Weihs and Magnum Chen and Marion Lepert and Marius Memmel and Masayoshi Tomizuka and Masha Itkina and Mateo Guaman Castro and Max Spero and Maximilian Du and Michael Ahn and Michael C. Yip and Mingtong Zhang and Mingyu Ding and Minho Heo and Mohan Kumar Srirama and Mohit Sharma and Moo Jin Kim and Muhammad Zubair Irshad and Naoaki Kanazawa and Nicklas Hansen and Nicolas Heess and Nikhil J Joshi and Niko Suenderhauf and Ning Liu and Norman Di Palo and Nur Muhammad Mahi Shafiullah and Oier Mees and Oliver Kroemer and Osbert Bastani and Pannag R Sanketi and Patrick \"Tree\" Miller and Patrick Yin and Paul Wohlhart and Peng Xu and Peter David Fagan and Peter Mitrano and Pierre Sermanet and Pieter Abbeel and Priya Sundaresan and Qiuyu Chen and Quan Vuong and Rafael Rafailov and Ran Tian and Ria Doshi and Roberto Martín-Martín and Rohan Baijal and Rosario Scalise and Rose Hendrix and Roy Lin and Runjia Qian and Ruohan Zhang and Russell Mendonca and Rutav Shah and Ryan Hoque and Ryan Julian and Samuel Bustamante and Sean Kirmani and Sergey Levine and Shan Lin and Sherry Moore and Shikhar Bahl and Shivin Dass and Shubham Sonawani and Shubham Tulsiani and Shuran Song and Sichun Xu and Siddhant Haldar and Siddharth Karamcheti and Simeon Adebola and Simon Guist and Soroush Nasiriany and Stefan Schaal and Stefan Welker and Stephen Tian and Subramanian Ramamoorthy and Sudeep Dasari and Suneel Belkhale and Sungjae Park and Suraj Nair and Suvir Mirchandani and Takayuki Osa and Tanmay Gupta and Tatsuya Harada and Tatsuya Matsushima and Ted Xiao and Thomas Kollar and Tianhe Yu and Tianli Ding and Todor Davchev and Tony Z. Zhao and Travis Armstrong and Trevor Darrell and Trinity Chung and Vidhi Jain and Vikash Kumar and Vincent Vanhoucke and Vitor Guizilini and Wei Zhan and Wenxuan Zhou and Wolfram Burgard and Xi Chen and Xiangyu Chen and Xiaolong Wang and Xinghao Zhu and Xinyang Geng and Xiyuan Liu and Xu Liangwei and Xuanlin Li and Yansong Pang and Yao Lu and Yecheng Jason Ma and Yejin Kim and Yevgen Chebotar and Yifan Zhou and Yifeng Zhu and Yilin Wu and Ying Xu and Yixuan Wang and Yonatan Bisk and Yongqiang Dou and Yoonyoung Cho and Youngwoon Lee and Yuchen Cui and Yue Cao and Yueh-Hua Wu and Yujin Tang and Yuke Zhu and Yunchu Zhang and Yunfan Jiang and Yunshuang Li and Yunzhu Li and Yusuke Iwasawa and Yutaka Matsuo and Zehan Ma and Zhuo Xu and Zichen Jeff Cui and Zichen Zhang and Zipeng Fu and Zipeng Lin},\n year = {2023},\n url = {https://doi.org/10.1109/ICRA57147.2024.10611477},\n doi = {10.1109/icra57147.2024.10611477}\n}", "preview": "assets/papers/open-x-embodiment.png", "previewSourceUrl": "https://arxiv.org/pdf/2310.08864" }, "metrics": { "citations": { "value": 110, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4387764390", "updatedAt": "2026-10-19", "recordTitle": "Open X-Embodiment: Robotic Learning Datasets and RT-X Models" }, "github": { "value": 2076, "source": "GitHub", "sourceUrl": "https://github.com/google-deepmind/open_x_embodiment/stargazers", "repository": "google-deepmind/open_x_embodiment", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "bridgedata-v2", "institutions": [ { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" }, { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "BridgeData V2", "title": "BridgeData V2: A Dataset for Robot Learning at Scale", "domains": [ "embodied" ], "kinds": [ "datasets", "papers" ], "year": 2023, "venue": "CoRL 2023", "summary": "Provides diverse low-cost robot manipulation trajectories across multiple environments for scalable imitation learning and offline reinforcement learning.", "tags": [ "Datasets", "Imitation learning" ], "environment": "WidowX manipulation across varied real-world environments.", "interface": "Image- and language-conditioned demonstration trajectories.", "limits": "Generalization depends on the chosen task, data split, and robot setup. Follow the source instructions for dataset access and preprocessing.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v229/walke23a.html", "data": "https://rail-berkeley.github.io/bridgedata/", "project": "https://rail-berkeley.github.io/bridgedata/", "code": "https://github.com/rail-berkeley/bridge_data_v2", "pdf": "https://proceedings.mlr.press/v229/walke23a/walke23a.pdf" }, "rank": 164, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "dataset", "credit": "Homer Rich Walke and collaborators", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v229/walke23a.html", "checkedAt": "2026-09-14", "authors": [ "Homer Rich Walke", "Kevin Black", "Tony Z. Zhao", "Quan Vuong", "Chongyi Zheng", "Philippe Hansen-Estruch", "Andre Wang He", "Vivek Myers", "Moo Jin Kim", "Max Du", "Abraham Lee", "Kuan Fang", "Chelsea Finn", "Sergey Levine" ], "date": "2023-12-02", "dateLabel": "Publisher record date", "bibtex": "@misc{bridgedata-v2,\n title = {BridgeData V2: A Dataset for Robot Learning at Scale},\n author = {Homer Rich Walke and Kevin Black and Tony Z. Zhao and Quan Vuong and Chongyi Zheng and Philippe Hansen-Estruch and Andre Wang He and Vivek Myers and Moo Jin Kim and Max Du and Abraham Lee and Kuan Fang and Chelsea Finn and Sergey Levine},\n year = {2023},\n url = {https://proceedings.mlr.press/v229/walke23a.html}\n}", "preview": "assets/papers/bridgedata-v2.png", "previewSourceUrl": "https://proceedings.mlr.press/v229/walke23a/walke23a.pdf" }, "metrics": { "citations": { "value": 12, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4386185624", "updatedAt": "2026-10-19", "recordTitle": "BridgeData V2: A Dataset for Robot Learning at Scale" }, "github": { "value": 292, "source": "GitHub", "sourceUrl": "https://github.com/rail-berkeley/bridge_data_v2/stargazers", "repository": "rail-berkeley/bridge_data_v2", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "droid", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "DROID", "title": "DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset", "domains": [ "embodied" ], "kinds": [ "datasets", "papers" ], "year": 2024, "venue": "RSS 2024", "summary": "Collects manipulation demonstrations across many scenes and institutions to study learning from diverse real-world experience.", "tags": [ "Datasets", "Imitation learning" ], "environment": "Distributed real-world manipulation with a shared robot setup.", "interface": "Multi-view observations, robot state, and demonstrated actions.", "limits": "Scene diversity does not imply support for arbitrary robots. Use the documented calibration, preprocessing, and data splits.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://www.roboticsproceedings.org/rss20/p120.html", "data": "https://droid-dataset.github.io/", "project": "https://droid-dataset.github.io/", "code": "https://github.com/droid-dataset/droid", "pdf": "https://www.roboticsproceedings.org/rss20/p120.pdf" }, "rank": 165, "selectionTrack": "peer-reviewed", "selectionReason": "Within the five-year publication window; official journal or conference record verified.", "paperType": "dataset", "credit": "Alexander Khazatsky and the DROID collaboration", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.roboticsproceedings.org/rss20/p120.html", "checkedAt": "2026-09-14", "authors": [ "Alexander Khazatsky", "Karl Pertsch", "Suraj Nair", "Ashwin Balakrishna", "Sudeep Dasari", "Siddharth Karamcheti", "Soroush Nasiriany", "Mohan Kumar Srirama", "Lawrence Yunliang Chen", "Kirsty Ellis", "Peter David Fagan", "Joey Hejna", "Masha Itkina", "Marion Lepert", "Yecheng Jason Ma", "Patrick Tree Miller", "Jimmy Wu", "Suneel Belkhale", "Shivin Dass", "Huy Ha", "Arhan Jain", "Abraham Lee", "Youngwoon Lee", "Marius Memmel", "Sungjae Park", "Ilija Radosavovic", "Kaiyuan Wang", "Albert Zhan", "Kevin Black", "Cheng Chi", "Kyle Beltran Hatch", "Shan Lin", "Jingpei Lu", "Jean Mercat", "Abdul Rehman", "Pannag R Sanketi", "Archit Sharma", "Cody Simpson", "Quan Vuong", "Homer Rich Walke", "Blake Wulfe", "Ted Xiao", "Jonathan Heewon Yang", "Arefeh Yavary", "Tony Z. Zhao", "Christopher Agia", "Rohan Baijal", "Mateo Guaman Castro", "Daphne Chen", "Qiuyu Chen", "Trinity Chung", "Jaimyn Drake", "Ethan Paul Foster", "Jensen Gao", "David Antonio Herrera", "Minho Heo", "Kyle Hsu", "Jiaheng Hu", "Donovon Jackson", "Charlotte Le", "Yunshuang Li", "Roy Lin", "Zehan Ma", "Abhiram Maddukuri", "Suvir Mirchandani", "Daniel Morton", "Tony Nguyen", "Abigail O'Neill", "Rosario Scalise", "Derick Seale", "Victor Son", "Stephen Tian", "Emi Tran", "Andrew E. Wang", "Yilin Wu", "Annie Xie", "Jingyun Yang", "Patrick Yin", "Yunchu Zhang", "Osbert Bastani", "Glen Berseth", "Jeannette Bohg", "Ken Goldberg", "Abhinav Gupta", "Abhishek Gupta", "Dinesh Jayaraman", "Joseph J Lim", "Jitendra Malik", "Roberto Martín-Martín", "Subramanian Ramamoorthy", "Dorsa Sadigh", "Shuran Song", "Jiajun Wu", "Michael C. Yip", "Yuke Zhu", "Thomas Kollar", "Sergey Levine", "Chelsea Finn" ], "date": "2024-07-15", "dateLabel": "Publisher record date", "bibtex": "@misc{droid,\n title = {DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset},\n author = {Alexander Khazatsky and Karl Pertsch and Suraj Nair and Ashwin Balakrishna and Sudeep Dasari and Siddharth Karamcheti and Soroush Nasiriany and Mohan Kumar Srirama and Lawrence Yunliang Chen and Kirsty Ellis and Peter David Fagan and Joey Hejna and Masha Itkina and Marion Lepert and Yecheng Jason Ma and Patrick Tree Miller and Jimmy Wu and Suneel Belkhale and Shivin Dass and Huy Ha and Arhan Jain and Abraham Lee and Youngwoon Lee and Marius Memmel and Sungjae Park and Ilija Radosavovic and Kaiyuan Wang and Albert Zhan and Kevin Black and Cheng Chi and Kyle Beltran Hatch and Shan Lin and Jingpei Lu and Jean Mercat and Abdul Rehman and Pannag R Sanketi and Archit Sharma and Cody Simpson and Quan Vuong and Homer Rich Walke and Blake Wulfe and Ted Xiao and Jonathan Heewon Yang and Arefeh Yavary and Tony Z. Zhao and Christopher Agia and Rohan Baijal and Mateo Guaman Castro and Daphne Chen and Qiuyu Chen and Trinity Chung and Jaimyn Drake and Ethan Paul Foster and Jensen Gao and David Antonio Herrera and Minho Heo and Kyle Hsu and Jiaheng Hu and Donovon Jackson and Charlotte Le and Yunshuang Li and Roy Lin and Zehan Ma and Abhiram Maddukuri and Suvir Mirchandani and Daniel Morton and Tony Nguyen and Abigail O'Neill and Rosario Scalise and Derick Seale and Victor Son and Stephen Tian and Emi Tran and Andrew E. Wang and Yilin Wu and Annie Xie and Jingyun Yang and Patrick Yin and Yunchu Zhang and Osbert Bastani and Glen Berseth and Jeannette Bohg and Ken Goldberg and Abhinav Gupta and Abhishek Gupta and Dinesh Jayaraman and Joseph J Lim and Jitendra Malik and Roberto Martín-Martín and Subramanian Ramamoorthy and Dorsa Sadigh and Shuran Song and Jiajun Wu and Michael C. Yip and Yuke Zhu and Thomas Kollar and Sergey Levine and Chelsea Finn},\n year = {2024},\n url = {https://www.roboticsproceedings.org/rss20/p120.html}\n}", "preview": "assets/papers/droid.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss20/p120.pdf" }, "metrics": { "citations": { "value": 151, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402354047", "updatedAt": "2026-10-19", "recordTitle": "DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset" }, "github": { "value": 457, "source": "GitHub", "sourceUrl": "https://github.com/droid-dataset/droid/stargazers", "repository": "droid-dataset/droid", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "article-vpt", "name": "Learning Minecraft from video", "title": "Learning to play Minecraft with Video PreTraining", "domains": [ "game" ], "kinds": [ "articles" ], "year": 2022, "venue": "OpenAI", "summary": "Explains how inverse-dynamics labels turn large video collections into training data for Minecraft behavior.", "tags": [ "Research perspectives" ], "environment": "Virtual environments.", "interface": "See the original article for the system interface and examples.", "limits": "Native keyboard and mouse control; separates pretraining from task fine-tuning.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://openai.com/index/vpt/" }, "rank": -20, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "OpenAI", "sourceDate": "2022-06-23", "relatedIds": [ "vpt" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "article-sima", "name": "SIMA: A generalist game agent", "title": "A generalist AI agent for 3D virtual environments", "domains": [ "game" ], "kinds": [ "articles" ], "year": 2024, "venue": "Google DeepMind", "summary": "Introduces a language-instructed agent trained across different virtual environments.", "tags": [ "Research perspectives" ], "environment": "Virtual environments.", "interface": "See the original article for the system interface and examples.", "limits": "Screen observations and keyboard/mouse actions; a research overview rather than a new benchmark record.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://deepmind.google/blog/sima-generalist-ai-agent-for-3d-virtual-environments/" }, "rank": -19, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "Google DeepMind", "sourceDate": "2024-03-13", "relatedIds": [ "sima2" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "article-sima2", "name": "SIMA 2: Playing, reasoning, and learning", "title": "SIMA 2: An Agent that Plays, Reasons, and Learns With You in Virtual 3D Worlds", "domains": [ "game" ], "kinds": [ "articles" ], "year": 2025, "venue": "Google DeepMind · SIMA Team", "summary": "Describes Gemini-based reasoning, interaction, and generalization in virtual worlds.", "tags": [ "Research perspectives" ], "environment": "Virtual environments.", "interface": "See the original article for the system interface and examples.", "limits": "Claims and examples concern virtual environments; they do not establish physical robot competence.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://deepmind.google/blog/sima-2-an-agent-that-plays-reasons-and-learns-with-you-in-virtual-3d-worlds/" }, "rank": -18, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "Google DeepMind · SIMA Team", "sourceDate": "2025-11-13", "relatedIds": [ "sima2" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "article-pi0", "name": "π0: Our first generalist policy", "title": "π0: Our First Generalist Policy", "domains": [ "embodied" ], "kinds": [ "articles" ], "year": 2024, "venue": "Physical Intelligence", "summary": "Introduces a vision-language-action policy with flow-matching action outputs and examples of adaptation to robot tasks.", "tags": [ "Research perspectives" ], "environment": "Robot learning and physical interaction.", "interface": "See the original article for the system interface and examples.", "limits": "Some demanding tasks use post-training; selected demonstrations are not general success guarantees.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://www.pi.website/blog/pi0" }, "rank": -17, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "Physical Intelligence", "sourceDate": "2024-10-31", "relatedIds": [ "openpi", "pi0-paper", "pi0-fast", "pi05", "pi05-ki", "pi05-ego", "pi07", "pi-star06" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-15" }, { "id": "article-lerobot", "name": "LeRobot: Imagine, evaluate, improve", "title": "LeRobot v0.6.0: Imagine, Evaluate, Improve", "domains": [ "embodied" ], "kinds": [ "articles" ], "year": 2026, "venue": "Hugging Face · LeRobot team", "summary": "Documents a robot-learning release connecting policy training, evaluation, reward models, and corrective deployment data.", "tags": [ "Research perspectives" ], "environment": "Robot learning and physical interaction.", "interface": "See the original article for the system interface and examples.", "limits": "Release-specific features require compatible hardware, dependencies, and model configuration.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://huggingface.co/blog/lerobot-release-v060" }, "rank": -16, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "Hugging Face · LeRobot team", "sourceDate": "2026-07-07", "relatedIds": [ "lerobot" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "article-robot-use", "name": "Robot-use agents", "title": "Robot-use agents", "domains": [ "embodied" ], "kinds": [ "articles" ], "year": 2026, "venue": "Phillip Isola · MIT", "summary": "Explores agents that reason over goals and operate robots through sensor and actuator interfaces.", "tags": [ "Research perspectives" ], "environment": "Robot learning and physical interaction.", "interface": "See the original article for the system interface and examples.", "limits": "A perspective on robot interfaces and agent loops, not a peer-reviewed evaluation.", "status": "selected", "evidence": "Official source checked; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "article": "https://web.mit.edu/phillipi/www/writing/robot-use-agents.html" }, "rank": -15, "selectionTrack": "research-article", "selectionReason": "Official author or research-team article providing context for the collection.", "credit": "Phillip Isola · MIT", "sourceDate": "2026-09-07", "relatedIds": [ "openvla", "lerobot" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-voyager-diamond", "name": "Voyager · Minecraft mining", "title": "Voyager · Minecraft mining", "domains": [ "game" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "A mining-skill example from the official Voyager Minecraft gallery.", "tags": [ "Game control" ], "environment": "Minecraft.", "interface": "JavaScript skill code, game state, and execution feedback.", "limits": "Programmatic skill execution with Mineflayer; the authors render demonstrations from recorded gameplay, which is not a record of pixel-only control.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://voyager.minedojo.org/", "video": "https://voyager.minedojo.org/assets/videos/gallery/diamond.mp4", "code": "https://github.com/MineDojo/Voyager" }, "rank": 172, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://voyager.minedojo.org/assets/videos/gallery/diamond.mp4", "type": "video/mp4" }, "demoType": "game-control", "credit": "Voyager research team", "relatedIds": [ "voyager" ], "metrics": { "github": { "value": 7271, "source": "GitHub", "sourceUrl": "https://github.com/MineDojo/Voyager/stargazers", "repository": "MineDojo/Voyager", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-voyager-cactus", "name": "Voyager · Collecting cactus", "title": "Voyager · Collecting cactus", "domains": [ "game" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "A second official Voyager example showing resource collection.", "tags": [ "Game control" ], "environment": "Minecraft.", "interface": "JavaScript skill code, game state, and execution feedback.", "limits": "A selected skill demonstration, not evidence of success on every task. The recording viewpoint is distinct from the programmatic observation interface.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://voyager.minedojo.org/", "video": "https://voyager.minedojo.org/assets/videos/gallery/cactus.mp4", "code": "https://github.com/MineDojo/Voyager" }, "rank": 173, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://voyager.minedojo.org/assets/videos/gallery/cactus.mp4", "type": "video/mp4" }, "demoType": "game-control", "credit": "Voyager research team", "relatedIds": [ "voyager" ], "metrics": { "github": { "value": 7271, "source": "GitHub", "sourceUrl": "https://github.com/MineDojo/Voyager/stargazers", "repository": "MineDojo/Voyager", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-gamengen", "name": "GameNGen · Neural DOOM", "title": "GameNGen · Neural DOOM", "domains": [ "game" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "A DOOM sequence generated by an action-conditioned neural simulator.", "tags": [ "World models" ], "environment": "DOOM / ViZDoom.", "interface": "Past game frames and player actions condition diffusion-based next-frame prediction. PPO primarily collects training trajectories.", "limits": "This demonstrates frame generation conditioned on actions. It is not an autonomous game-playing policy or a measured combat result.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://gamengen.github.io/", "video": "https://gamengen.github.io/static/videos/e1m1_t.mp4" }, "rank": 174, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://gamengen.github.io/static/videos/e1m1_t.mp4", "type": "video/mp4" }, "demoType": "generated-world", "credit": "GameNGen research team", "relatedIds": [ "kb-w01" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-diamond", "name": "DIAMOND · Generated CS:GO", "title": "DIAMOND · Generated CS:GO", "domains": [ "game" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "An interactive Counter-Strike sequence from the diffusion world-model project.", "tags": [ "World models" ], "environment": "Atari; Counter-Strike: Global Offensive, Dust II.", "interface": "Human CS:GO frames and actions feed low-resolution dynamics and super-resolution generation. The shooter extension does not train an RL policy.", "limits": "Keep the CS:GO generation extension separate from Atari policy-learning results. Visual continuity is not a combat success metric.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://diamond-wm.github.io/", "video": "https://diamond-wm.github.io/static/videos/8.mp4", "code": "https://github.com/eloialonso/diamond" }, "rank": 175, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://diamond-wm.github.io/static/videos/8.mp4", "type": "video/mp4" }, "demoType": "generated-world", "credit": "DIAMOND research team", "relatedIds": [ "kb-w02" ], "metrics": { "github": { "value": 2122, "source": "GitHub", "sourceUrl": "https://github.com/eloialonso/diamond/stargazers", "repository": "eloialonso/diamond", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-openvla", "name": "OpenVLA · Manipulation in clutter", "title": "OpenVLA · Manipulation in clutter", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "A real WidowX manipulation rollout from the OpenVLA project gallery.", "tags": [ "Robot learning" ], "environment": "Multiple real-world robot manipulation platforms.", "interface": "RGB images and language map to robot actions.", "limits": "The source clip is labeled 1.5× speed. This selected qualitative rollout is not a control-frequency measurement or an aggregate success rate.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://openvla.github.io/", "video": "https://openvla.github.io/static/videos/carousel/bridge_pick_clutter_2.mp4", "code": "https://github.com/openvla/openvla" }, "rank": 176, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://openvla.github.io/static/videos/carousel/bridge_pick_clutter_2.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "OpenVLA research team", "relatedIds": [ "openvla" ], "metrics": { "github": { "value": 7171, "source": "GitHub", "sourceUrl": "https://github.com/openvla/openvla/stargazers", "repository": "openvla/openvla", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-aloha", "name": "ACT / ALOHA · Slotting a battery", "title": "ACT / ALOHA · Slotting a battery", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "A learned ACT policy performs a precise bimanual battery-insertion task.", "tags": [ "Robot learning" ], "environment": "The ALOHA bimanual platform.", "interface": "Images and joint observations map to action chunks.", "limits": "The source labels this a real-time learned-policy rollout, separate from its teleoperation examples. This selected task does not establish arbitrary dexterity.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://tonyzhaozh.github.io/aloha/", "video": "https://tonyzhaozh.github.io/aloha/resources/slot_battery.mp4" }, "rank": 177, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://tonyzhaozh.github.io/aloha/resources/slot_battery.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "Tony Zhao and the ALOHA / ACT team", "relatedIds": [ "act" ], "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-diffusion-mug", "name": "Diffusion Policy · Mug manipulation", "title": "Diffusion Policy · Mug manipulation", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "An official real-robot manipulation example using a diffusion action policy.", "tags": [ "Robot learning" ], "environment": "Simulated and real-world robot manipulation tasks.", "interface": "Visual observations map to continuous action sequences.", "limits": "A selected physical execution example; the action distribution, demonstration data, and evaluation setup remain task-specific.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://diffusion-policy.cs.columbia.edu/", "video": "https://diffusion-policy.cs.columbia.edu/videos/highlight_mug.mp4", "code": "https://github.com/real-stanford/diffusion_policy" }, "rank": 178, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://diffusion-policy.cs.columbia.edu/videos/highlight_mug.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "Cheng Chi and the Diffusion Policy team", "relatedIds": [ "diffusion-policy" ], "metrics": { "github": { "value": 4656, "source": "GitHub", "sourceUrl": "https://github.com/real-stanford/diffusion_policy/stargazers", "repository": "real-stanford/diffusion_policy", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-diffusion-lift", "name": "Diffusion Policy · Simulated lifting", "title": "Diffusion Policy · Simulated lifting", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "A simulation rollout from the diffusion-policy evaluation gallery.", "tags": [ "Simulation" ], "environment": "Simulated and real-world robot manipulation tasks.", "interface": "Visual observations map to continuous action sequences.", "limits": "Simulation evidence is distinct from the real-robot videos. A short successful rollout is not an aggregate success rate.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://diffusion-policy.cs.columbia.edu/", "video": "https://diffusion-policy.cs.columbia.edu/videos/lift.mp4", "code": "https://github.com/real-stanford/diffusion_policy" }, "rank": 179, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related research source.", "video": { "src": "https://diffusion-policy.cs.columbia.edu/videos/lift.mp4", "type": "video/mp4" }, "demoType": "simulation", "credit": "Cheng Chi and the Diffusion Policy team", "relatedIds": [ "diffusion-policy" ], "metrics": { "github": { "value": 4656, "source": "GitHub", "sourceUrl": "https://github.com/real-stanford/diffusion_policy/stargazers", "repository": "real-stanford/diffusion_policy", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "efficientzero", "institutions": [ { "name": "Tsinghua University", "url": "https://www.tsinghua.edu.cn/en/", "logo": "assets/institutions/tsinghua.png", "mark": "THU" }, { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" } ], "name": "EfficientZero", "title": "Mastering Atari Games with Limited Data", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2021, "venue": "NeurIPS 2021", "summary": "Adds value-prefix prediction and consistency learning to MuZero, reaching strong Atari 100k performance from roughly two hours of experience.", "tags": [ "Atari", "Model-based RL" ], "environment": "Atari 100k and DeepMind Control Suite.", "interface": "Pixels and rewards feed a learned dynamics model; Monte Carlo tree search selects discrete actions.", "limits": "Atari 100k is a fixed low-data protocol. Reported normalized scores do not imply transfer to unseen games or robust real-time play.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper/2021/hash/d5eca8dc3820cad9fe56a3bafda65ca1-Abstract.html", "code": "https://github.com/YeWR/EfficientZero", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2021/file/d5eca8dc3820cad9fe56a3bafda65ca1-Paper.pdf" }, "rank": 31, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper/2021/hash/d5eca8dc3820cad9fe56a3bafda65ca1-Abstract.html", "checkedAt": "2026-09-14", "authors": [ "Weirui Ye", "Shaohuai Liu", "Thanard Kurutach", "Pieter Abbeel", "Yang Gao" ], "date": "2021-12-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{efficientzero,\n title = {Mastering Atari Games with Limited Data},\n author = {Weirui Ye and Shaohuai Liu and Thanard Kurutach and Pieter Abbeel and Yang Gao},\n booktitle = {NeurIPS 2021},\n year = {2021},\n url = {https://proceedings.neurips.cc/paper/2021/hash/d5eca8dc3820cad9fe56a3bafda65ca1-Abstract.html}\n}", "preview": "assets/papers/efficientzero.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2021/file/d5eca8dc3820cad9fe56a3bafda65ca1-Paper.pdf" }, "metrics": { "citations": { "value": 65, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3208460365", "updatedAt": "2026-10-19", "recordTitle": "Mastering Atari Games with Limited Data" }, "github": { "value": 944, "source": "GitHub", "sourceUrl": "https://github.com/YeWR/EfficientZero/stargazers", "repository": "YeWR/EfficientZero", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14", "publicationDate": "2021-12-06" }, { "id": "gato", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "Gato", "title": "A Generalist Agent", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "TMLR 2022", "summary": "Tokenizes observations and actions from hundreds of tasks so one Transformer can play games, caption images, converse, and control a robot arm.", "tags": [ "Generalist agents", "Multi-game" ], "environment": "Atari, simulated control, vision-language tasks, dialogue, and real robot block stacking.", "interface": "A shared token sequence represents observations, text, actions, and task context.", "limits": "Breadth spans many task formats, but performance varies widely and the real-robot evaluation is narrow.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=1ikK0kHjvj", "project": "https://deepmind.google/discover/blog/a-generalist-agent/", "pdf": "https://export.arxiv.org/pdf/2205.06175" }, "rank": 32, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://openreview.net/forum?id=1ikK0kHjvj", "checkedAt": "2026-09-14", "authors": [ "Scott Reed", "Konrad Żołna", "Emilio Parisotto", "Sergio Gómez Colmenarejo", "Alexander Novikov", "Gabriel Barth-Maron", "Mai Giménez", "Yury Sulsky", "Jackie Kay", "Jost Tobias Springenberg", "Tom Eccles", "Jake Bruce", "Ali Razavi", "Ashley Edwards", "Nicolas Heess", "Yutian Chen", "Raia Hadsell", "Oriol Vinyals", "Mahyar Bordbar", "Nando de Freitas" ], "date": "2022-11-10", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{gato,\n title = {A Generalist Agent},\n author = {Scott Reed and Konrad Żołna and Emilio Parisotto and Sergio Gómez Colmenarejo and Alexander Novikov and Gabriel Barth-Maron and Mai Giménez and Yury Sulsky and Jackie Kay and Jost Tobias Springenberg and Tom Eccles and Jake Bruce and Ali Razavi and Ashley Edwards and Nicolas Heess and Yutian Chen and Raia Hadsell and Oriol Vinyals and Mahyar Bordbar and Nando de Freitas},\n booktitle = {TMLR 2022},\n year = {2022},\n url = {https://openreview.net/forum?id=1ikK0kHjvj}\n}", "preview": "assets/papers/gato.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2205.06175" }, "metrics": { "citations": { "value": 63, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4307868887", "updatedAt": "2026-10-19", "recordTitle": "A Generalist Agent" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "cicero", "institutions": [ { "name": "Meta AI", "url": "https://ai.meta.com/", "mark": "META" } ], "name": "CICERO", "title": "Human-level play in the game of Diplomacy by combining language models with strategic reasoning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "Science 2022", "summary": "Combines strategic planning with controllable dialogue to negotiate, coordinate, and compete with people in full-press Diplomacy.", "tags": [ "Strategy games", "Language agents" ], "environment": "Seven-player online Diplomacy with natural-language negotiation.", "interface": "Game state, dialogue history, strategic intents, and legal orders are coupled through planning and language modules.", "limits": "Human-level results come from a defined online league and speed-game format; honesty controls reduce but do not eliminate all undesirable dialogue.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://doi.org/10.1126/science.ade9097", "project": "https://ai.meta.com/research/cicero/", "code": "https://github.com/facebookresearch/diplomacy_cicero", "pdf": "https://arxiv.org/pdf/2211.08966" }, "rank": 33, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://doi.org/10.1126/science.ade9097", "checkedAt": "2026-09-14", "authors": [ "Anton Bakhtin", "Noam Brown", "Emily Dinan", "Gabriele Farina", "Colin Flaherty", "Daniel Fried", "Andrew Goff", "Jonathan Gray", "Hengyuan Hu", "Athul Paul Jacob", "Mojtaba Komeili", "Karthik Konath", "Minae Kwon", "Adam Lerer", "Mike Lewis", "Alexander Miller", "Sasha Mitts", "Adithya Renduchintala", "Stephen Roller", "Dirk Rowe", "Weiyan Shi", "Joe Spisak", "Alexander Wei", "David Wu", "Hugh Zhang", "Markus Zijlstra" ], "date": "2022-11-22", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{cicero,\n title = {Human-level play in the game of Diplomacy by combining language models with strategic reasoning},\n author = {Anton Bakhtin and Noam Brown and Emily Dinan and Gabriele Farina and Colin Flaherty and Daniel Fried and Andrew Goff and Jonathan Gray and Hengyuan Hu and Athul Paul Jacob and Mojtaba Komeili and Karthik Konath and Minae Kwon and Adam Lerer and Mike Lewis and Alexander Miller and Sasha Mitts and Adithya Renduchintala and Stephen Roller and Dirk Rowe and Weiyan Shi and Joe Spisak and Alexander Wei and David Wu and Hugh Zhang and Markus Zijlstra},\n booktitle = {Science 2022},\n year = {2022},\n url = {https://doi.org/10.1126/science.ade9097}\n}", "preview": "assets/papers/cicero.png", "previewSourceUrl": "https://arxiv.org/pdf/2211.08966" }, "metrics": { "citations": { "value": 236, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4309663019", "updatedAt": "2026-10-19", "recordTitle": "Human-level play in the game of Diplomacy by combining language models with strategic reasoning" }, "github": { "value": 1432, "source": "GitHub", "sourceUrl": "https://github.com/facebookresearch/diplomacy_cicero/stargazers", "repository": "facebookresearch/diplomacy_cicero", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "deepnash", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "DeepNash", "title": "Mastering the game of Stratego with model-free multiagent reinforcement learning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "Science 2022", "summary": "Uses regularized Nash dynamics and model-free self-play to reach expert-level Stratego without search during play.", "tags": [ "Strategy games", "Multi-agent RL" ], "environment": "Stratego, a long-horizon two-player game with hidden information.", "interface": "Board observations and legal moves drive a model-free policy trained entirely through self-play.", "limits": "The result is specialized to one ruleset and platform; expert ranking does not establish competence in unrelated strategic settings.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://doi.org/10.1126/science.add4679", "project": "https://deepmind.google/discover/blog/mastering-stratego-the-classic-game-of-imperfect-information/", "pdf": "https://arxiv.org/pdf/2206.15378" }, "rank": 34, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://doi.org/10.1126/science.add4679", "checkedAt": "2026-09-14", "authors": [ "Julien Pérolat", "Bart De Vylder", "Daniel Hennes", "Eugene Tarassov", "Florian Strub", "Vincent C. J. de Boer", "Paul Müller", "Jerome T. Connor", "Neil Burch", "Thomas Anthony", "Stephen McAleer", "Romuald Élie", "Sarah H. Cen", "Zhe Wang", "Audrūnas Gruslys", "Aleksandra Malysheva", "Mina Khan", "Sherjil Ozair", "Finbarr Timbers", "Toby Pohlen", "Tom Eccles", "Mark Rowland", "Marc Lanctot", "Jean-Baptiste Lespiau", "Bilal Piot", "Shayegan Omidshafiei", "Edward Lockhart", "Laurent Sifre", "Nathalie Beauguerlange", "Rémi Munos", "David Silver", "Satinder Singh", "Demis Hassabis", "Karl Tuyls" ], "date": "2022-12-01", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{deepnash,\n title = {Mastering the game of Stratego with model-free multiagent reinforcement learning},\n author = {Julien Pérolat and Bart De Vylder and Daniel Hennes and Eugene Tarassov and Florian Strub and Vincent C. J. de Boer and Paul Müller and Jerome T. Connor and Neil Burch and Thomas Anthony and Stephen McAleer and Romuald Élie and Sarah H. Cen and Zhe Wang and Audrūnas Gruslys and Aleksandra Malysheva and Mina Khan and Sherjil Ozair and Finbarr Timbers and Toby Pohlen and Tom Eccles and Mark Rowland and Marc Lanctot and Jean-Baptiste Lespiau and Bilal Piot and Shayegan Omidshafiei and Edward Lockhart and Laurent Sifre and Nathalie Beauguerlange and Rémi Munos and David Silver and Satinder Singh and Demis Hassabis and Karl Tuyls},\n booktitle = {Science 2022},\n year = {2022},\n url = {https://doi.org/10.1126/science.add4679}\n}", "preview": "assets/papers/deepnash.png", "previewSourceUrl": "https://arxiv.org/pdf/2206.15378" }, "metrics": { "citations": { "value": 154, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4310705874", "updatedAt": "2026-09-28", "recordTitle": "Mastering the game of Stratego with model-free multiagent reinforcement learning" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "iris", "institutions": [ { "name": "University of Geneva", "url": "https://www.unige.ch/en/", "mark": "UNIGE" } ], "name": "IRIS", "title": "Transformers are Sample-Efficient World Models", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICLR 2023", "summary": "Learns a discrete visual tokenizer and autoregressive Transformer world model, then trains an agent through imagined Atari rollouts.", "tags": [ "World models", "Atari" ], "environment": "Atari 100k across 26 games.", "interface": "Image tokens, actions, and rewards form autoregressive sequences for imagined policy learning.", "limits": "Long-rollout fidelity and game coverage remain bounded by the training distribution; Atari scores are protocol-specific.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=vhFu1Acb0xb", "code": "https://github.com/eloialonso/iris", "pdf": "https://export.arxiv.org/pdf/2209.00588" }, "rank": 35, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://openreview.net/forum?id=vhFu1Acb0xb", "checkedAt": "2026-09-14", "authors": [ "Vincent Micheli", "Eloi Alonso", "François Fleuret" ], "date": "2023-05-01", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{iris,\n title = {Transformers are Sample-Efficient World Models},\n author = {Vincent Micheli and Eloi Alonso and François Fleuret},\n booktitle = {ICLR 2023},\n year = {2023},\n url = {https://openreview.net/forum?id=vhFu1Acb0xb}\n}", "preview": "assets/papers/iris.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2209.00588" }, "metrics": { "citations": { "value": 25, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4294435908", "updatedAt": "2026-10-19", "recordTitle": "Transformers are Sample-Efficient World Models" }, "github": { "value": 902, "source": "GitHub", "sourceUrl": "https://github.com/eloialonso/iris/stargazers", "repository": "eloialonso/iris", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "genie", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "Genie", "title": "Genie: Generative Interactive Environments", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICML 2024", "summary": "Learns latent actions and video dynamics from unlabeled gameplay, turning a single image or prompt into an interactable generated environment.", "tags": [ "Generative environments", "World models" ], "environment": "Platformer videos and generated 2D interactive scenes.", "interface": "An initial frame and inferred latent action sequence condition an autoregressive video model.", "limits": "Short-horizon generated interaction can drift and does not reproduce a full game engine, ruleset, or persistent world state.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v235/bruce24a.html", "project": "https://sites.google.com/view/genie-2024/", "pdf": "https://export.arxiv.org/pdf/2402.15391" }, "rank": 36, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v235/bruce24a.html", "checkedAt": "2026-09-14", "authors": [ "Jake Bruce", "Michael Dennis", "Ashley Edwards", "Jack Parker-Holder", "Yuge Shi", "Edward Hughes", "Matthew Lai", "Aditi Mavalankar", "Richie Steigerwald", "Chris Apps", "Yusuf Aytar", "Sarah Bechtle", "Feryal Behbahani", "Stephanie Chan", "Nicolas Heess", "Lucy González", "Simon Osindero", "Sherjil Ozair", "Scott Reed", "Jingwei Zhang", "Konrad Żołna", "Jeff Clune", "Nando de Freitas", "Satinder Singh", "Tim Rocktäschel" ], "date": "2024-05-02", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{genie,\n title = {Genie: Generative Interactive Environments},\n author = {Jake Bruce and Michael Dennis and Ashley Edwards and Jack Parker-Holder and Yuge Shi and Edward Hughes and Matthew Lai and Aditi Mavalankar and Richie Steigerwald and Chris Apps and Yusuf Aytar and Sarah Bechtle and Feryal Behbahani and Stephanie Chan and Nicolas Heess and Lucy González and Simon Osindero and Sherjil Ozair and Scott Reed and Jingwei Zhang and Konrad Żołna and Jeff Clune and Nando de Freitas and Satinder Singh and Tim Rocktäschel},\n booktitle = {ICML 2024},\n year = {2024},\n url = {https://proceedings.mlr.press/v235/bruce24a.html}\n}", "preview": "assets/papers/genie.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2402.15391" }, "metrics": { "citations": { "value": 12, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4392182423", "updatedAt": "2026-09-21", "recordTitle": "Genie: Generative Interactive Environments" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "bedd", "institutions": [ { "name": "Carnegie Mellon University", "url": "https://www.cmu.edu/", "mark": "CMU" } ], "name": "BEDD", "title": "BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Formalizes four fuzzy Minecraft tasks with demonstrations and dense human preferences for training and evaluating agents from human feedback.", "tags": [ "Minecraft", "Human feedback" ], "environment": "MineRL BASALT tasks such as building a village house or creating a waterfall.", "interface": "Gameplay videos, image-action pairs, pairwise human judgments, and natural-language rationales.", "limits": "The benchmark captures four subjective tasks and fixed comparison data; leaderboard results depend on the human-evaluation protocol.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/67a6726dcd555b982cabb3446ffac01d-Abstract.html", "code": "https://github.com/minerllabs/basalt-benchmark", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/67a6726dcd555b982cabb3446ffac01d-Paper-Datasets_and_Benchmarks.pdf" }, "rank": 37, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/67a6726dcd555b982cabb3446ffac01d-Abstract.html", "checkedAt": "2026-09-14", "authors": [ "Stephanie Milani", "Anssi Kanervisto", "Karolis Ramanauskas", "Sander Schulhoff", "Brandon Houghton", "Rohin Shah" ], "date": "2023-12-05", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{bedd,\n title = {BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks},\n author = {Stephanie Milani and Anssi Kanervisto and Karolis Ramanauskas and Sander Schulhoff and Brandon Houghton and Rohin Shah},\n booktitle = {NeurIPS 2023},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/67a6726dcd555b982cabb3446ffac01d-Abstract.html}\n}", "preview": "assets/papers/bedd.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/67a6726dcd555b982cabb3446ffac01d-Paper-Datasets_and_Benchmarks.pdf" }, "metrics": { "citations": { "value": 1, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4389469779", "updatedAt": "2026-10-12", "recordTitle": "BEDD: The MineRL BASALT Evaluation and Demonstrations Dataset for Training and Benchmarking Agents that Solve Fuzzy Tasks" }, "github": { "value": 22, "source": "GitHub", "sourceUrl": "https://github.com/minerllabs/basalt-benchmark/stargazers", "repository": "minerllabs/basalt-benchmark", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "crafter", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "Crafter", "title": "Benchmarking the Spectrum of Agent Capabilities", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "ICLR 2022", "summary": "Introduces a fast open-world survival game whose achievement spectrum measures exploration, generalization, memory, and long-horizon control.", "tags": [ "Open-world", "Benchmarks" ], "environment": "Procedurally generated 2D survival worlds with resources, crafting, hazards, and achievements.", "interface": "Pixels, discrete actions, rewards, and semantically named achievements.", "limits": "Crafter is deliberately compact and stylized; broad achievement coverage is not equivalent to transfer across commercial games.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=1W0z96MFEoH", "project": "https://danijar.com/project/crafter/", "code": "https://github.com/danijar/crafter", "pdf": "https://export.arxiv.org/pdf/2109.06780" }, "rank": 38, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "benchmark", "publicationDate": "2022-04-25", "bibliography": { "source": "Publisher", "sourceUrl": "https://openreview.net/forum?id=1W0z96MFEoH", "checkedAt": "2026-09-14", "authors": [ "Danijar Hafner" ], "date": "2022-04-25", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{crafter,\n title = {Benchmarking the Spectrum of Agent Capabilities},\n author = {Danijar Hafner},\n booktitle = {ICLR 2022},\n year = {2022},\n url = {https://openreview.net/forum?id=1W0z96MFEoH}\n}", "preview": "assets/papers/crafter.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2109.06780" }, "metrics": { "citations": { "value": 16, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3201036908", "updatedAt": "2026-10-19", "recordTitle": "Benchmarking the Spectrum of Agent Capabilities" }, "github": { "value": 604, "source": "GitHub", "sourceUrl": "https://github.com/danijar/crafter/stargazers", "repository": "danijar/crafter", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "multi-game-dt", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "Multi-Game Decision Transformer", "title": "Multi-Game Decision Transformers", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2022, "venue": "NeurIPS 2022", "summary": "Scales return-conditioned sequence modeling to one offline Transformer that plays up to 46 Atari games and adapts through fine-tuning.", "tags": [ "Offline RL", "Multi-game" ], "environment": "Forty-six Atari games from fixed offline replay datasets.", "interface": "Returns, observations, actions, rewards, and game identity are modeled as token sequences.", "limits": "Training is offline and Atari-specific; close-to-human aggregate performance hides substantial variation by game.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/b2cac94f82928a85055987d9fd44753f-Abstract-Conference.html", "project": "https://sites.google.com/view/multi-game-transformers", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2022/file/b2cac94f82928a85055987d9fd44753f-Paper-Conference.pdf" }, "rank": 39, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/hash/b2cac94f82928a85055987d9fd44753f-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Kuang-Huei Lee", "Ofir Nachum", "Mengjiao Yang", "Lisa Lee", "Daniel Freeman", "Winnie Xu", "Sergio Guadarrama", "Ian Fischer", "Eric Jang", "Henryk Michalewski", "Igor Mordatch" ], "date": "2022-12-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{multi_game_dt,\n title = {Multi-Game Decision Transformers},\n author = {Kuang-Huei Lee and Ofir Nachum and Mengjiao Yang and Lisa Lee and Daniel Freeman and Winnie Xu and Sergio Guadarrama and Ian Fischer and Eric Jang and Henryk Michalewski and Igor Mordatch},\n booktitle = {NeurIPS 2022},\n year = {2022},\n url = {https://proceedings.neurips.cc/paper_files/paper/2022/hash/b2cac94f82928a85055987d9fd44753f-Abstract-Conference.html}\n}", "preview": "assets/papers/multi-game-dt.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2022/file/b2cac94f82928a85055987d9fd44753f-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 27, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4281663178", "updatedAt": "2026-09-15", "recordTitle": "Multi-Game Decision Transformers" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "diplodocus", "institutions": [ { "name": "Meta AI", "url": "https://ai.meta.com/", "mark": "META" } ], "name": "Diplodocus", "title": "Mastering the Game of No-Press Diplomacy via Human-Regularized Reinforcement Learning and Planning", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICLR 2023", "summary": "Regularizes self-play toward human conventions so planning agents can cooperate and compete effectively with people in no-press Diplomacy.", "tags": [ "Strategy games", "Human coordination" ], "environment": "Seven-player no-press Diplomacy tournaments with human participants.", "interface": "Game states and legal orders feed a human-regularized planning and reinforcement-learning system.", "limits": "The no-press variant removes natural-language negotiation; tournament results depend on the participant pool and evaluation format.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=F61FwJTZhb", "project": "https://web.mit.edu/~gfarina/www/2023/no_press_diplomacy_iclr23/", "code": "https://github.com/facebookresearch/diplomacy_cicero", "pdf": "https://export.arxiv.org/pdf/2210.05492" }, "rank": 40, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://openreview.net/forum?id=F61FwJTZhb", "checkedAt": "2026-09-14", "authors": [ "Anton Bakhtin", "David J. Wu", "Adam Lerer", "Jonathan Gray", "Athul Paul Jacob", "Gabriele Farina", "Alexander H. Miller", "Noam Brown" ], "date": "2023-05-01", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{diplodocus,\n title = {Mastering the Game of No-Press Diplomacy via Human-Regularized Reinforcement Learning and Planning},\n author = {Anton Bakhtin and David J. Wu and Adam Lerer and Jonathan Gray and Athul Paul Jacob and Gabriele Farina and Alexander H. Miller and Noam Brown},\n booktitle = {ICLR 2023},\n year = {2023},\n url = {https://openreview.net/forum?id=F61FwJTZhb}\n}", "preview": "assets/papers/diplodocus.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2210.05492" }, "metrics": { "citations": { "value": 12, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4305033123", "updatedAt": "2026-10-19", "recordTitle": "Mastering the Game of No-Press Diplomacy via Human-Regularized Reinforcement Learning and Planning" }, "github": { "value": 1432, "source": "GitHub", "sourceUrl": "https://github.com/facebookresearch/diplomacy_cicero/stargazers", "repository": "facebookresearch/diplomacy_cicero", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "decision-transformer", "institutions": [ { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" }, { "name": "Meta AI", "url": "https://ai.meta.com/", "mark": "META" } ], "name": "Decision Transformer", "title": "Decision Transformer: Reinforcement Learning via Sequence Modeling", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2021, "venue": "NeurIPS 2021", "summary": "Recasts offline reinforcement learning as autoregressive sequence modeling conditioned on desired return, including Atari control.", "tags": [ "Offline RL", "Sequence modeling" ], "environment": "Atari, D4RL locomotion, and Key-to-Door tasks.", "interface": "Desired returns, states, and previous actions are interleaved as Transformer tokens.", "limits": "Performance depends on dataset coverage and chosen return targets; the model does not explore or improve from new online experience by itself.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2021/hash/7f489f642a0ddb10272b5c31057f0663-Abstract.html", "project": "https://sites.google.com/berkeley.edu/decision-transformer", "code": "https://github.com/kzl/decision-transformer", "pdf": "https://proceedings.neurips.cc/paper/2021/file/7f489f642a0ddb10272b5c31057f0663-Paper.pdf" }, "rank": 41, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "publicationDate": "2021-12-06", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2021/hash/7f489f642a0ddb10272b5c31057f0663-Abstract.html", "checkedAt": "2026-09-14", "authors": [ "Lili Chen", "Kevin Lu", "Aravind Rajeswaran", "Kimin Lee", "Aditya Grover", "Misha Laskin", "Pieter Abbeel", "Aravind Srinivas", "Igor Mordatch" ], "date": "2021-12-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{decision_transformer,\n title = {Decision Transformer: Reinforcement Learning via Sequence Modeling},\n author = {Lili Chen and Kevin Lu and Aravind Rajeswaran and Kimin Lee and Aditya Grover and Misha Laskin and Pieter Abbeel and Aravind Srinivas and Igor Mordatch},\n booktitle = {NeurIPS 2021},\n year = {2021},\n url = {https://proceedings.neurips.cc/paper_files/paper/2021/hash/7f489f642a0ddb10272b5c31057f0663-Abstract.html}\n}", "preview": "assets/papers/decision-transformer.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper/2021/file/7f489f642a0ddb10272b5c31057f0663-Paper.pdf" }, "metrics": { "citations": { "value": 459, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3169291081", "updatedAt": "2026-10-12", "recordTitle": "Decision Transformer: Reinforcement Learning via Sequence Modeling" }, "github": { "value": 2842, "source": "GitHub", "sourceUrl": "https://github.com/kzl/decision-transformer/stargazers", "repository": "kzl/decision-transformer", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "lightzero", "institutions": [ { "name": "Shanghai AI Laboratory", "url": "https://www.shlab.org.cn/", "mark": "SHLAB" } ], "name": "LightZero", "title": "LightZero: A Unified Benchmark for Monte Carlo Tree Search in General Sequential Decision Scenarios", "domains": [ "game" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Unifies MuZero-style tree-search systems and evaluates them across board games, Atari, continuous control, MiniGrid, and GoBigger.", "tags": [ "Tree search", "Benchmarks" ], "environment": "Atari, board games, MuJoCo, MiniGrid, and GoBigger.", "interface": "Modular learned models, policies, value estimates, and Monte Carlo tree search.", "limits": "A common software framework does not make scores across heterogeneous tasks directly comparable; many settings still require task-specific tuning.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/765043fe026f7d704c96cec027f13843-Abstract-Datasets_and_Benchmarks.html", "code": "https://github.com/opendilab/LightZero", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/765043fe026f7d704c96cec027f13843-Paper-Datasets_and_Benchmarks.pdf" }, "rank": 42, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/765043fe026f7d704c96cec027f13843-Abstract-Datasets_and_Benchmarks.html", "checkedAt": "2026-09-14", "authors": [ "Yazhe Niu", "Yuan Pu", "Zhenjie Yang", "Xueyan Li", "Tong Zhou", "Jiyuan Ren", "Shuai Hu", "Hongsheng Li", "Yu Liu" ], "date": "2023-12-10", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{lightzero,\n title = {LightZero: A Unified Benchmark for Monte Carlo Tree Search in General Sequential Decision Scenarios},\n author = {Yazhe Niu and Yuan Pu and Zhenjie Yang and Xueyan Li and Tong Zhou and Jiyuan Ren and Shuai Hu and Hongsheng Li and Yu Liu},\n booktitle = {NeurIPS 2023},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/765043fe026f7d704c96cec027f13843-Abstract-Datasets_and_Benchmarks.html}\n}", "preview": "assets/papers/lightzero.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/765043fe026f7d704c96cec027f13843-Paper-Datasets_and_Benchmarks.pdf" }, "metrics": { "citations": { "value": 2, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4387635629", "updatedAt": "2026-10-12", "recordTitle": "LightZero: A Unified Benchmark for Monte Carlo Tree Search in General Sequential Decision Scenarios" }, "github": { "value": 1652, "source": "GitHub", "sourceUrl": "https://github.com/opendilab/LightZero/stargazers", "repository": "opendilab/LightZero", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "saycan", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "SayCan", "title": "Do As I Can, Not As I Say: Grounding Language in Robotic Affordances", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2022, "venue": "CoRL 2022", "summary": "Scores language-model plans by the robot skills that are actually feasible, grounding long-horizon instructions in learned affordances.", "tags": [ "Language planning", "Affordances" ], "environment": "Real mobile manipulation in office and kitchen-like spaces.", "interface": "Natural-language instructions are decomposed into skills and reweighted by learned value functions.", "limits": "The planner is bounded by a fixed skill library and affordance models; semantic plausibility does not guarantee safe execution.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v205/ichter23a.html", "project": "https://say-can.github.io/", "pdf": "https://proceedings.mlr.press/v205/ichter23a/ichter23a.pdf" }, "rank": 43, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v205/ichter23a.html", "checkedAt": "2026-09-14", "authors": [ "Michael Ahn", "Anthony Brohan", "Noah Brown", "Yevgen Chebotar", "Omar Cortes", "Byron David", "Chelsea Finn", "Chuyuan Fu", "Keerthana Gopalakrishnan", "Karol Hausman", "Alex Herzog", "Daniel Ho", "Jasmine Hsu", "Julian Ibarz", "Brian Ichter", "Alex Irpan", "Eric Jang", "Rosario Jauregui Ruano", "Kyle Jeffrey", "Sally Jesmonth", "Nikhil Joshi", "Ryan Julian", "Dmitry Kalashnikov", "Yuheng Kuang", "Kuang-Huei Lee", "Sergey Levine", "Yao Lu", "Linda Luu", "Carolina Parada", "Peter Pastor", "Jornell Quiambao", "Kanishka Rao", "Jarek Rettinghouse", "Diego Reyes", "Pierre Sermanet", "Nicolas Sievers", "Clayton Tan", "Alexander Toshev", "Vincent Vanhoucke", "Fei Xia", "Ted Xiao", "Peng Xu", "Sichun Xu", "Mengyuan Yan", "Andy Zeng" ], "date": "2023-03-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{saycan,\n title = {Do As I Can, Not As I Say: Grounding Language in Robotic Affordances},\n author = {Michael Ahn and Anthony Brohan and Noah Brown and Yevgen Chebotar and Omar Cortes and Byron David and Chelsea Finn and Chuyuan Fu and Keerthana Gopalakrishnan and Karol Hausman and Alex Herzog and Daniel Ho and Jasmine Hsu and Julian Ibarz and Brian Ichter and Alex Irpan and Eric Jang and Rosario Jauregui Ruano and Kyle Jeffrey and Sally Jesmonth and Nikhil Joshi and Ryan Julian and Dmitry Kalashnikov and Yuheng Kuang and Kuang-Huei Lee and Sergey Levine and Yao Lu and Linda Luu and Carolina Parada and Peter Pastor and Jornell Quiambao and Kanishka Rao and Jarek Rettinghouse and Diego Reyes and Pierre Sermanet and Nicolas Sievers and Clayton Tan and Alexander Toshev and Vincent Vanhoucke and Fei Xia and Ted Xiao and Peng Xu and Sichun Xu and Mengyuan Yan and Andy Zeng},\n booktitle = {CoRL 2022},\n year = {2022},\n url = {https://proceedings.mlr.press/v205/ichter23a.html}\n}", "preview": "assets/papers/saycan.png", "previewSourceUrl": "https://proceedings.mlr.press/v205/ichter23a/ichter23a.pdf" }, "metrics": { "citations": { "value": 523, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4224912544", "updatedAt": "2026-09-15", "recordTitle": "Do As I Can, Not As I Say: Grounding Language in Robotic Affordances" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "palm-e", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "PaLM-E", "title": "PaLM-E: An Embodied Multimodal Language Model", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICML 2023", "summary": "Injects visual and continuous sensor embeddings into a large language model for robot planning, visual reasoning, and transfer across embodiments.", "tags": [ "Embodied language models", "Multimodal reasoning" ], "environment": "Mobile manipulation, tabletop control, visual question answering, and task-and-motion planning.", "interface": "Text, images, robot state, and neural scene representations are interleaved in one language-model context.", "limits": "PaLM-E primarily produces plans and representations; low-level execution relies on separate controllers and task-specific systems.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v202/driess23a.html", "project": "https://palm-e.github.io/", "pdf": "https://proceedings.mlr.press/v202/driess23a/driess23a.pdf" }, "rank": 44, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v202/driess23a.html", "checkedAt": "2026-09-14", "authors": [ "Danny Driess", "Fei Xia", "Mehdi S. M. Sajjadi", "Corey Lynch", "Aakanksha Chowdhery", "Brian Ichter", "Ayzaan Wahid", "Jonathan Tompson", "Quan Vuong", "Tianhe Yu", "Wenlong Huang", "Yevgen Chebotar", "Pierre Sermanet", "Daniel Duckworth", "Sergey Levine", "Vincent Vanhoucke", "Karol Hausman", "Marc Toussaint", "Klaus Greff", "Andy Zeng", "Igor Mordatch", "Pete Florence" ], "date": "2023-07-03", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{palm_e,\n title = {PaLM-E: An Embodied Multimodal Language Model},\n author = {Danny Driess and Fei Xia and Mehdi S. M. Sajjadi and Corey Lynch and Aakanksha Chowdhery and Brian Ichter and Ayzaan Wahid and Jonathan Tompson and Quan Vuong and Tianhe Yu and Wenlong Huang and Yevgen Chebotar and Pierre Sermanet and Daniel Duckworth and Sergey Levine and Vincent Vanhoucke and Karol Hausman and Marc Toussaint and Klaus Greff and Andy Zeng and Igor Mordatch and Pete Florence},\n booktitle = {ICML 2023},\n year = {2023},\n url = {https://proceedings.mlr.press/v202/driess23a.html}\n}", "preview": "assets/papers/palm-e.png", "previewSourceUrl": "https://proceedings.mlr.press/v202/driess23a/driess23a.pdf" }, "metrics": { "citations": { "value": 357, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4323572061", "updatedAt": "2026-09-21", "recordTitle": "PaLM-E: An Embodied Multimodal Language Model" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "rt1", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" }, { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "RT-1", "title": "RT-1: Robotics Transformer for Real-World Control at Scale", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "RSS 2023", "summary": "Trains an efficient Transformer policy on 130,000 real-robot episodes and more than 700 language-conditioned tasks.", "tags": [ "Robot transformers", "Multi-task learning" ], "environment": "A fleet of Everyday Robots performing mobile-manipulation tasks in real office kitchens.", "interface": "Camera images and language instructions map directly to discretized arm, gripper, and base actions.", "limits": "The dataset uses a related robot fleet and controlled workspaces; deployment to new hardware or safety-critical settings requires adaptation.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://roboticsproceedings.org/rss19/p025.html", "project": "https://robotics-transformer.github.io/", "code": "https://github.com/google-research/robotics_transformer", "pdf": "https://www.roboticsproceedings.org/rss19/p025.pdf" }, "rank": 45, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://roboticsproceedings.org/rss19/p025.html", "checkedAt": "2026-09-14", "authors": [ "Anthony Brohan", "Noah Brown", "Justice Carbajal", "Yevgen Chebotar", "Joseph Dabis", "Chelsea Finn", "Keerthana Gopalakrishnan", "Karol Hausman", "Alexander Herzog", "Jasmine Hsu", "Julian Ibarz", "Brian Ichter", "Alex Irpan", "Tomas Jackson", "Sally Jesmonth", "Nikhil Joshi", "Ryan Julian", "Dmitry Kalashnikov", "Yuheng Kuang", "Isabel Leal", "Kuang-Huei Lee", "Sergey Levine", "Yao Lu", "Utsav Malla", "Deeksha Manjunath", "Igor Mordatch", "Ofir Nachum", "Carolina Parada", "Jodilyn Peralta", "Emily Perez", "Karl Pertsch", "Jornell Quiambao", "Kanishka Rao", "Michael Ryoo", "Grecia Salazar", "Pannag Sanketi", "Kevin Sayed", "Jaspiar Singh", "Sumedh Sontakke", "Austin Stone", "Clayton Tan", "Huong Tran", "Vincent Vanhoucke", "Steve Vega", "Quan Vuong", "Fei Xia", "Ted Xiao", "Peng Xu", "Sichun Xu", "Tianhe Yu", "Brianna Zitkovich" ], "date": "2023-07-10", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{rt1,\n title = {RT-1: Robotics Transformer for Real-World Control at Scale},\n author = {Anthony Brohan and Noah Brown and Justice Carbajal and Yevgen Chebotar and Joseph Dabis and Chelsea Finn and Keerthana Gopalakrishnan and Karol Hausman and Alexander Herzog and Jasmine Hsu and Julian Ibarz and Brian Ichter and Alex Irpan and Tomas Jackson and Sally Jesmonth and Nikhil Joshi and Ryan Julian and Dmitry Kalashnikov and Yuheng Kuang and Isabel Leal and Kuang-Huei Lee and Sergey Levine and Yao Lu and Utsav Malla and Deeksha Manjunath and Igor Mordatch and Ofir Nachum and Carolina Parada and Jodilyn Peralta and Emily Perez and Karl Pertsch and Jornell Quiambao and Kanishka Rao and Michael Ryoo and Grecia Salazar and Pannag Sanketi and Kevin Sayed and Jaspiar Singh and Sumedh Sontakke and Austin Stone and Clayton Tan and Huong Tran and Vincent Vanhoucke and Steve Vega and Quan Vuong and Fei Xia and Ted Xiao and Peng Xu and Sichun Xu and Tianhe Yu and Brianna Zitkovich},\n booktitle = {RSS 2023},\n year = {2023},\n url = {https://roboticsproceedings.org/rss19/p025.html}\n}", "preview": "assets/papers/rt1.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss19/p025.pdf" }, "metrics": { "citations": { "value": 707, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4385430679", "updatedAt": "2026-10-19", "recordTitle": "RT-1: Robotics Transformer for Real-World Control at Scale" }, "github": { "value": 1741, "source": "GitHub", "sourceUrl": "https://github.com/google-research/robotics_transformer/stargazers", "repository": "google-research/robotics_transformer", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "vima", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" }, { "name": "NVIDIA", "url": "https://www.nvidia.com/", "logo": "assets/institutions/nvidia.png", "mark": "NV" }, { "name": "Tsinghua University", "url": "https://www.tsinghua.edu.cn/en/", "logo": "assets/institutions/tsinghua.png", "mark": "THU" } ], "name": "VIMA", "title": "VIMA: General Robot Manipulation with Multimodal Prompts", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICML 2023", "summary": "Expresses many simulated manipulation tasks as prompts that interleave language, images, and demonstrations, then predicts robot actions autoregressively.", "tags": [ "Multimodal prompting", "Manipulation" ], "environment": "VIMA-Bench with procedurally generated tabletop manipulation tasks.", "interface": "Text and visual prompt tokens condition a Transformer that emits object-centric motor actions.", "limits": "Results are primarily in simulation with privileged object representations; real-world perception and control introduce additional error sources.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v202/jiang23b.html", "project": "https://vimalabs.github.io/", "code": "https://github.com/vimalabs/VIMA", "pdf": "https://proceedings.mlr.press/v202/jiang23b/jiang23b.pdf" }, "rank": 46, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v202/jiang23b.html", "checkedAt": "2026-09-14", "authors": [ "Yunfan Jiang", "Agrim Gupta", "Zichen Zhang", "Guanzhi Wang", "Yongqiang Dou", "Yanjun Chen", "Li Fei-Fei", "Anima Anandkumar", "Yuke Zhu", "Linxi Fan" ], "date": "2023-07-23", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{vima,\n title = {VIMA: General Robot Manipulation with Multimodal Prompts},\n author = {Yunfan Jiang and Agrim Gupta and Zichen Zhang and Guanzhi Wang and Yongqiang Dou and Yanjun Chen and Li Fei-Fei and Anima Anandkumar and Yuke Zhu and Linxi Fan},\n booktitle = {ICML 2023},\n year = {2023},\n url = {https://proceedings.mlr.press/v202/jiang23b.html}\n}", "preview": "assets/papers/vima.png", "previewSourceUrl": "https://proceedings.mlr.press/v202/jiang23b/jiang23b.pdf" }, "metrics": { "citations": { "value": 94, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4303648971", "updatedAt": "2026-10-19", "recordTitle": "VIMA: General Robot Manipulation with Multimodal Prompts" }, "github": { "value": 858, "source": "GitHub", "sourceUrl": "https://github.com/vimalabs/VIMA/stargazers", "repository": "vimalabs/VIMA", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "peract", "institutions": [ { "name": "University of Washington", "url": "https://www.washington.edu/", "mark": "UW" }, { "name": "NVIDIA", "url": "https://www.nvidia.com/", "logo": "assets/institutions/nvidia.png", "mark": "NV" } ], "name": "PerAct", "title": "Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2022, "venue": "CoRL 2022", "summary": "Uses a Perceiver over voxelized RGB-D observations to detect language-conditioned 6-DoF actions for many tasks with one policy.", "tags": [ "3D policies", "Language conditioning" ], "environment": "RLBench simulation and seven real-world tabletop tasks.", "interface": "Language and RGB-D voxel grids map to discretized translation, rotation, gripper, and collision actions.", "limits": "Voxelization is memory-intensive and relies on calibrated depth; performance depends on keyframe extraction and task demonstrations.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v205/shridhar23a.html", "project": "https://peract.github.io/", "code": "https://github.com/peract/peract", "pdf": "https://proceedings.mlr.press/v205/shridhar23a/shridhar23a.pdf" }, "rank": 47, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v205/shridhar23a.html", "checkedAt": "2026-09-14", "authors": [ "Mohit Shridhar", "Lucas Manuelli", "Dieter Fox" ], "date": "2023-03-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{peract,\n title = {Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation},\n author = {Mohit Shridhar and Lucas Manuelli and Dieter Fox},\n booktitle = {CoRL 2022},\n year = {2022},\n url = {https://proceedings.mlr.press/v205/shridhar23a.html}\n}", "preview": "assets/papers/peract.png", "previewSourceUrl": "https://proceedings.mlr.press/v205/shridhar23a/shridhar23a.pdf" }, "metrics": { "citations": { "value": 128, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4297812239", "updatedAt": "2026-10-19", "recordTitle": "Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation" }, "github": { "value": 497, "source": "GitHub", "sourceUrl": "https://github.com/peract/peract/stargazers", "repository": "peract/peract", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "r3m", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" }, { "name": "Meta AI", "url": "https://ai.meta.com/", "mark": "META" } ], "name": "R3M", "title": "R3M: A Universal Visual Representation for Robot Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2022, "venue": "CoRL 2022", "summary": "Pretrains a reusable robot visual encoder on Ego4D human video through temporal, language, and sparsity objectives.", "tags": [ "Visual pretraining", "Robot learning" ], "environment": "Twelve simulated tasks and real Franka manipulation in a cluttered apartment.", "interface": "A frozen video-pretrained encoder supplies compact visual features to downstream imitation or reinforcement-learning policies.", "limits": "A strong average representation is not universally best; adaptation and downstream policy design still determine task success.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v205/nair23a.html", "project": "https://sites.google.com/view/robot-r3m/", "code": "https://github.com/facebookresearch/r3m", "pdf": "https://proceedings.mlr.press/v205/nair23a/nair23a.pdf" }, "rank": 48, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v205/nair23a.html", "checkedAt": "2026-09-14", "authors": [ "Suraj Nair", "Aravind Rajeswaran", "Vikash Kumar", "Chelsea Finn", "Abhinav Gupta" ], "date": "2023-03-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{r3m,\n title = {R3M: A Universal Visual Representation for Robot Manipulation},\n author = {Suraj Nair and Aravind Rajeswaran and Vikash Kumar and Chelsea Finn and Abhinav Gupta},\n booktitle = {CoRL 2022},\n year = {2022},\n url = {https://proceedings.mlr.press/v205/nair23a.html}\n}", "preview": "assets/papers/r3m.png", "previewSourceUrl": "https://proceedings.mlr.press/v205/nair23a/nair23a.pdf" }, "metrics": { "citations": { "value": 80, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4221159977", "updatedAt": "2026-10-19", "recordTitle": "R3M: A Universal Visual Representation for Robot Manipulation" }, "github": { "value": 376, "source": "GitHub", "sourceUrl": "https://github.com/facebookresearch/r3m/stargazers", "repository": "facebookresearch/r3m", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "vc1", "institutions": [ { "name": "Meta AI", "url": "https://ai.meta.com/", "mark": "META" }, { "name": "University of California, Berkeley", "url": "https://www.berkeley.edu/", "mark": "UCB" }, { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "VC-1", "title": "Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence?", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "NeurIPS 2023", "summary": "Builds CortexBench and VC-1 to compare large pretrained visual representations across navigation, locomotion, and manipulation.", "tags": [ "Embodied vision", "Benchmarks" ], "environment": "Seventeen simulated and real embodied tasks spanning navigation, locomotion, dexterity, and mobile manipulation.", "interface": "Frozen or adapted visual encoders feed task-specific policies across heterogeneous embodiments.", "limits": "No representation wins every task; averages across CortexBench should not be treated as a universal policy score.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/022ca1bed6b574b962c48a2856eb207b-Abstract-Conference.html", "project": "https://eai-vc.github.io/", "code": "https://github.com/facebookresearch/eai-vc", "pdf": "https://proceedings.neurips.cc/paper_files/paper/2023/file/022ca1bed6b574b962c48a2856eb207b-Paper-Conference.pdf" }, "rank": 49, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "benchmark", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/hash/022ca1bed6b574b962c48a2856eb207b-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Arjun Majumdar", "Karmesh Yadav", "Sergio Arnaud", "Jason Ma", "Claire Chen", "Sneha Silwal", "Aryan Jain", "Vincent-Pierre Berges", "Tingfan Wu", "Jay Vakil", "Pieter Abbeel", "Jitendra Malik", "Dhruv Batra", "Yixin Lin", "Oleksandr Maksymets", "Aravind Rajeswaran", "Franziska Meier" ], "date": "2023-12-10", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{vc1,\n title = {Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence?},\n author = {Arjun Majumdar and Karmesh Yadav and Sergio Arnaud and Jason Ma and Claire Chen and Sneha Silwal and Aryan Jain and Vincent-Pierre Berges and Tingfan Wu and Jay Vakil and Pieter Abbeel and Jitendra Malik and Dhruv Batra and Yixin Lin and Oleksandr Maksymets and Aravind Rajeswaran and Franziska Meier},\n booktitle = {NeurIPS 2023},\n year = {2023},\n url = {https://proceedings.neurips.cc/paper_files/paper/2023/hash/022ca1bed6b574b962c48a2856eb207b-Abstract-Conference.html}\n}", "preview": "assets/papers/vc1.png", "previewSourceUrl": "https://proceedings.neurips.cc/paper_files/paper/2023/file/022ca1bed6b574b962c48a2856eb207b-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 21, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4362514550", "updatedAt": "2026-10-05", "recordTitle": "Where are we in the search for an Artificial Visual Cortex for Embodied Intelligence?" }, "github": { "value": 509, "source": "GitHub", "sourceUrl": "https://github.com/facebookresearch/eai-vc/stargazers", "repository": "facebookresearch/eai-vc", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "robocat", "institutions": [ { "name": "Google DeepMind", "url": "https://deepmind.google/", "logo": "assets/institutions/google-deepmind.png", "mark": "DM" } ], "name": "RoboCat", "title": "RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "TMLR 2024", "summary": "Trains one goal-conditioned Transformer across robot arms, then uses its own successful rollouts to expand data for new tasks and embodiments.", "tags": [ "Generalist policies", "Self-improvement" ], "environment": "Multiple simulated and real robot arms with heterogeneous observation and action spaces.", "interface": "Visual goals, observations, and embodiment-specific action tokens are modeled as one sequence.", "limits": "Self-generated data is filtered and task-scoped; adaptation still needs hundreds of target examples and controlled evaluation.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://openreview.net/forum?id=vsCpLiIWHu", "project": "https://deepmind.google/discover/blog/robocat-a-self-improving-robotic-agent/", "pdf": "https://export.arxiv.org/pdf/2306.11706" }, "rank": 50, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://openreview.net/forum?id=vsCpLiIWHu", "checkedAt": "2026-09-14", "authors": [ "Konstantinos Bousmalis", "Giulia Vezzani", "Dushyant Rao", "Coline Devin", "Alex Lee", "Maria Bauzá", "Todor Davchev", "Yuxiang Zhou", "Agrim Gupta", "Akhil Raju", "Antoine Laurens", "Claudio Fantacci", "Valentin Dalibard", "Martina Zambelli", "Murilo Martins", "Rugile Pevceviciute", "Michiel Blokzijl", "Misha Denil", "Nathan Batchelor", "Thomas Lampe", "Emilio Parisotto", "Konrad Żołna", "Scott Reed", "Sergio Gómez Colmenarejo", "Jon Scholz", "Abbas Abdolmaleki", "Oliver Groth", "Jean-Baptiste Regli", "Oleg Sushkov", "Tom Rothörl", "José Enrique Chen", "Yusuf Aytar", "Dave Barker", "Joy Ortiz", "Martin Riedmiller", "Jost Tobias Springenberg", "Raia Hadsell", "Francesco Nori", "Nicolas Heess" ], "date": "2024-02-07", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{robocat,\n title = {RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation},\n author = {Konstantinos Bousmalis and Giulia Vezzani and Dushyant Rao and Coline Devin and Alex Lee and Maria Bauzá and Todor Davchev and Yuxiang Zhou and Agrim Gupta and Akhil Raju and Antoine Laurens and Claudio Fantacci and Valentin Dalibard and Martina Zambelli and Murilo Martins and Rugile Pevceviciute and Michiel Blokzijl and Misha Denil and Nathan Batchelor and Thomas Lampe and Emilio Parisotto and Konrad Żołna and Scott Reed and Sergio Gómez Colmenarejo and Jon Scholz and Abbas Abdolmaleki and Oliver Groth and Jean-Baptiste Regli and Oleg Sushkov and Tom Rothörl and José Enrique Chen and Yusuf Aytar and Dave Barker and Joy Ortiz and Martin Riedmiller and Jost Tobias Springenberg and Raia Hadsell and Francesco Nori and Nicolas Heess},\n booktitle = {TMLR 2024},\n year = {2024},\n url = {https://openreview.net/forum?id=vsCpLiIWHu}\n}", "preview": "assets/papers/robocat.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2306.11706" }, "metrics": { "citations": { "value": 9, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4381586832", "updatedAt": "2026-10-12", "recordTitle": "RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "mobile-aloha", "institutions": [ { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "Mobile ALOHA", "title": "Mobile ALOHA: Learning Bimanual Mobile Manipulation using Low-Cost Whole-Body Teleoperation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "CoRL 2024", "summary": "Extends ALOHA with a mobile base and whole-body teleoperation, enabling imitation of long-horizon bimanual household tasks.", "tags": [ "Mobile manipulation", "Bimanual control" ], "environment": "Real kitchens, elevators, and household spaces with a mobile dual-arm robot.", "interface": "Whole-body teleoperation records synchronized base, arm, gripper, and camera trajectories for behavior cloning.", "limits": "Tasks use curated demonstrations and a specific low-cost platform; success rates are sensitive to scene setup and co-training data.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v270/fu25b.html", "project": "https://mobile-aloha.github.io/", "code": "https://github.com/MarkFzp/mobile-aloha", "pdf": "https://export.arxiv.org/pdf/2401.02117" }, "rank": 51, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v270/fu25b.html", "checkedAt": "2026-09-14", "authors": [ "Zipeng Fu", "Tony Z. Zhao", "Chelsea Finn" ], "date": "2025-01-12", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{mobile_aloha,\n title = {Mobile ALOHA: Learning Bimanual Mobile Manipulation using Low-Cost Whole-Body Teleoperation},\n author = {Zipeng Fu and Tony Z. Zhao and Chelsea Finn},\n booktitle = {CoRL 2024},\n year = {2024},\n url = {https://proceedings.mlr.press/v270/fu25b.html}\n}", "preview": "assets/papers/mobile-aloha.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2401.02117" }, "metrics": { "citations": { "value": 31, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4390635563", "updatedAt": "2026-09-14", "recordTitle": "Mobile ALOHA: Learning Bimanual Mobile Manipulation using Low-Cost Whole-Body Teleoperation" }, "github": { "value": 4474, "source": "GitHub", "sourceUrl": "https://github.com/MarkFzp/mobile-aloha/stargazers", "repository": "MarkFzp/mobile-aloha", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "gr1", "institutions": [ { "name": "ByteDance Research", "url": "https://www.bytedance.com/en/", "mark": "BD" } ], "name": "GR-1", "title": "Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICLR 2024", "summary": "Pretrains one GPT-style model to predict video, then fine-tunes it to emit actions and future images for multi-task robot manipulation.", "tags": [ "Video pretraining", "Generative policies" ], "environment": "CALVIN simulation and real tabletop manipulation.", "interface": "Language, image history, and robot state tokens condition future-image and action prediction.", "limits": "Reported generalization covers defined scenes and objects; deployment remains tied to the fine-tuned embodiment and observation setup.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/2c37c5bcef24b9541550261dcd63261b-Abstract-Conference.html", "project": "https://gr1-manipulation.github.io/", "code": "https://github.com/bytedance/GR-1", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2024/file/2c37c5bcef24b9541550261dcd63261b-Paper-Conference.pdf" }, "rank": 52, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/2c37c5bcef24b9541550261dcd63261b-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Hongtao Wu", "Ya Jing", "Chilam Cheang", "Guangzeng Chen", "Jiafeng Xu", "Xinghang Li", "Minghuan Liu", "Hang Li", "Tao Kong" ], "date": "2024-05-07", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{gr1,\n title = {Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation},\n author = {Hongtao Wu and Ya Jing and Chilam Cheang and Guangzeng Chen and Jiafeng Xu and Xinghang Li and Minghuan Liu and Hang Li and Tao Kong},\n booktitle = {ICLR 2024},\n year = {2024},\n url = {https://proceedings.iclr.cc/paper_files/paper/2024/hash/2c37c5bcef24b9541550261dcd63261b-Abstract-Conference.html}\n}", "preview": "assets/papers/gr1.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/file/2c37c5bcef24b9541550261dcd63261b-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 4, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4390136276", "updatedAt": "2026-09-21", "recordTitle": "Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation" }, "github": { "value": 308, "source": "GitHub", "sourceUrl": "https://github.com/bytedance/GR-1/stargazers", "repository": "bytedance/GR-1", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "3d-diffuser-actor", "institutions": [ { "name": "Carnegie Mellon University", "url": "https://www.cmu.edu/", "mark": "CMU" } ], "name": "3D Diffuser Actor", "title": "3D Diffuser Actor: Policy Diffusion with 3D Scene Representations", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "CoRL 2024", "summary": "Denoises 3D end-effector trajectories while attending to scene geometry, language, and proprioception for multi-task manipulation.", "tags": [ "3D policies", "Diffusion models" ], "environment": "RLBench, CALVIN, and a small set of real-robot manipulation tasks.", "interface": "Multi-view RGB-D features, language, and proprioception condition a 3D trajectory diffusion Transformer.", "limits": "Performance assumes depth and camera calibration; real-world evidence uses a limited number of demonstrations and tasks.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v270/ke25a.html", "project": "https://3d-diffuser-actor.github.io/", "pdf": "https://export.arxiv.org/pdf/2402.10885" }, "rank": 53, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v270/ke25a.html", "checkedAt": "2026-09-14", "authors": [ "Tsung-Wei Ke", "Nikolaos Gkanatsios", "Katerina Fragkiadaki" ], "date": "2025-01-12", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{3d_diffuser_actor,\n title = {3D Diffuser Actor: Policy Diffusion with 3D Scene Representations},\n author = {Tsung-Wei Ke and Nikolaos Gkanatsios and Katerina Fragkiadaki},\n booktitle = {CoRL 2024},\n year = {2024},\n url = {https://proceedings.mlr.press/v270/ke25a.html}\n}", "preview": "assets/papers/3d-diffuser-actor.png", "previewSourceUrl": "https://export.arxiv.org/pdf/2402.10885" }, "metrics": { "citations": { "value": 4, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4391949089", "updatedAt": "2026-10-19", "recordTitle": "3D Diffuser Actor: Policy Diffusion with 3D Scene Representations" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "dp3", "institutions": [ { "name": "Tsinghua University", "url": "https://www.tsinghua.edu.cn/en/", "logo": "assets/institutions/tsinghua.png", "mark": "THU" }, { "name": "Shanghai AI Laboratory", "url": "https://www.shlab.org.cn/", "mark": "SHLAB" } ], "name": "DP3", "title": "3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "RSS 2024", "summary": "Pairs compact point-cloud encodings with diffusion policies to improve low-demo learning and spatial generalization across dexterous tasks.", "tags": [ "3D policies", "Imitation learning" ], "environment": "Seventy-two simulated tasks and four real-robot manipulation tasks.", "interface": "Sparse point clouds and proprioception condition a diffusion model over action sequences.", "limits": "Depth sensing and point-cloud preprocessing are required; broad simulation counts include heterogeneous tasks with different difficulty.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://www.roboticsproceedings.org/rss20/p067.html", "project": "https://3d-diffusion-policy.github.io/", "code": "https://github.com/YanjieZe/3D-Diffusion-Policy", "pdf": "https://www.roboticsproceedings.org/rss20/p067.pdf" }, "rank": 54, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://www.roboticsproceedings.org/rss20/p067.html", "checkedAt": "2026-09-14", "authors": [ "Yanjie Ze", "Gu Zhang", "Kangning Zhang", "Chenyuan Hu", "Muhan Wang", "Huazhe Xu" ], "date": "2024-07-15", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{dp3,\n title = {3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations},\n author = {Yanjie Ze and Gu Zhang and Kangning Zhang and Chenyuan Hu and Muhan Wang and Huazhe Xu},\n booktitle = {RSS 2024},\n year = {2024},\n url = {https://www.roboticsproceedings.org/rss20/p067.html}\n}", "preview": "assets/papers/dp3.png", "previewSourceUrl": "https://www.roboticsproceedings.org/rss20/p067.pdf" }, "metrics": { "citations": { "value": 220, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4402354045", "updatedAt": "2026-10-19", "recordTitle": "3D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representations" }, "github": { "value": 1461, "source": "GitHub", "sourceUrl": "https://github.com/YanjieZe/3D-Diffusion-Policy/stargazers", "repository": "YanjieZe/3D-Diffusion-Policy", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "cliport", "institutions": [ { "name": "University of Washington", "url": "https://www.washington.edu/", "mark": "UW" }, { "name": "NVIDIA", "url": "https://www.nvidia.com/", "logo": "assets/institutions/nvidia.png", "mark": "NV" } ], "name": "CLIPort", "title": "CLIPort: What and Where Pathways for Robotic Manipulation", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2021, "venue": "CoRL 2021", "summary": "Combines CLIP semantics with Transporter-style spatial precision for language-conditioned tabletop manipulation.", "tags": [ "Vision-language", "Manipulation" ], "environment": "Ten simulated and nine real tabletop tasks including packing, sorting, and cloth folding.", "interface": "RGB observations and language instructions feed separate semantic and spatial streams that predict pick-and-place actions.", "limits": "The action abstraction is specialized to tabletop manipulation and does not cover arbitrary trajectories or force-sensitive skills.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v164/shridhar22a.html", "project": "https://cliport.github.io/", "code": "https://github.com/cliport/cliport", "pdf": "https://proceedings.mlr.press/v164/shridhar22a/shridhar22a.pdf" }, "rank": 55, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "publicationDate": "2021-11-08", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v164/shridhar22a.html", "checkedAt": "2026-09-14", "authors": [ "Mohit Shridhar", "Lucas Manuelli", "Dieter Fox" ], "date": "2022-01-11", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{cliport,\n title = {CLIPort: What and Where Pathways for Robotic Manipulation},\n author = {Mohit Shridhar and Lucas Manuelli and Dieter Fox},\n booktitle = {CoRL 2021},\n year = {2021},\n url = {https://proceedings.mlr.press/v164/shridhar22a.html}\n}", "preview": "assets/papers/cliport.png", "previewSourceUrl": "https://proceedings.mlr.press/v164/shridhar22a/shridhar22a.pdf" }, "metrics": { "citations": { "value": 223, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W3203511201", "updatedAt": "2026-10-19", "recordTitle": "CLIPort: What and Where Pathways for Robotic Manipulation" }, "github": { "value": 547, "source": "GitHub", "sourceUrl": "https://github.com/cliport/cliport/stargazers", "repository": "cliport/cliport", "updatedAt": "2026-10-19" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "inner-monologue", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "Inner Monologue", "title": "Inner Monologue: Embodied Reasoning through Planning with Language Models", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2022, "venue": "CoRL 2022", "summary": "Feeds success detectors, scene descriptions, and human feedback back into an LLM planner so robot plans can react to execution outcomes.", "tags": [ "Language planning", "Closed-loop reasoning" ], "environment": "Simulated and real tabletop rearrangement plus long-horizon mobile manipulation.", "interface": "Natural-language instructions and textual environment feedback update high-level plans over a skill library.", "limits": "Feedback quality and the available skills bound recovery; language-level replanning does not solve low-level control failures.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.mlr.press/v205/huang23c.html", "project": "https://innermonologue.github.io/", "pdf": "https://proceedings.mlr.press/v205/huang23c/huang23c.pdf" }, "rank": 56, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.mlr.press/v205/huang23c.html", "checkedAt": "2026-09-14", "authors": [ "Wenlong Huang", "Fei Xia", "Ted Xiao", "Harris Chan", "Jacky Liang", "Pete Florence", "Andy Zeng", "Jonathan Tompson", "Igor Mordatch", "Yevgen Chebotar", "Pierre Sermanet", "Noah Brown", "Tomas Jackson", "Linda Luu", "Sergey Levine", "Karol Hausman", "Brian Ichter" ], "date": "2023-03-06", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{inner_monologue,\n title = {Inner Monologue: Embodied Reasoning through Planning with Language Models},\n author = {Wenlong Huang and Fei Xia and Ted Xiao and Harris Chan and Jacky Liang and Pete Florence and Andy Zeng and Jonathan Tompson and Igor Mordatch and Yevgen Chebotar and Pierre Sermanet and Noah Brown and Tomas Jackson and Linda Luu and Sergey Levine and Karol Hausman and Brian Ichter},\n booktitle = {CoRL 2022},\n year = {2022},\n url = {https://proceedings.mlr.press/v205/huang23c.html}\n}", "preview": "assets/papers/inner-monologue.png", "previewSourceUrl": "https://proceedings.mlr.press/v205/huang23c/huang23c.pdf" }, "metrics": { "citations": { "value": 208, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4285428875", "updatedAt": "2026-09-15", "recordTitle": "Inner Monologue: Embodied Reasoning through Planning with Language Models" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "rt-trajectory", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" }, { "name": "Stanford University", "url": "https://www.stanford.edu/", "logo": "assets/institutions/stanford.png", "mark": "SU" } ], "name": "RT-Trajectory", "title": "RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2024, "venue": "ICLR 2024", "summary": "Conditions robot policies on rough trajectory sketches that can come from drawings, videos, or generators, enabling new motion-centric tasks.", "tags": [ "Task conditioning", "Generalization" ], "environment": "Large-scale real-robot manipulation with novel task and object combinations.", "interface": "Camera observations and coarse 2D trajectory sketches condition low-level robot actions.", "limits": "Sketches provide motion guidance but still require compatible training data and perception; arbitrary new skills are not guaranteed.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/c383e44d9a878d1982d9abb838bd5d8a-Abstract-Conference.html", "project": "https://rt-trajectory.github.io/", "pdf": "https://proceedings.iclr.cc/paper_files/paper/2024/file/c383e44d9a878d1982d9abb838bd5d8a-Paper-Conference.pdf" }, "rank": 57, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/hash/c383e44d9a878d1982d9abb838bd5d8a-Abstract-Conference.html", "checkedAt": "2026-09-14", "authors": [ "Jiayuan Gu", "Sean Kirmani", "Paul Wohlhart", "Yao Lu", "Montserrat Gonzalez Arenas", "Kanishka Rao", "Wenhao Yu", "Chuyuan Fu", "Keerthana Gopalakrishnan", "Zhuo Xu", "Priya Sundaresan", "Peng Xu", "Hao Su", "Karol Hausman", "Chelsea Finn", "Quan Vuong", "Ted Xiao" ], "date": "2024-05-07", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{rt_trajectory,\n title = {RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches},\n author = {Jiayuan Gu and Sean Kirmani and Paul Wohlhart and Yao Lu and Montserrat Gonzalez Arenas and Kanishka Rao and Wenhao Yu and Chuyuan Fu and Keerthana Gopalakrishnan and Zhuo Xu and Priya Sundaresan and Peng Xu and Hao Su and Karol Hausman and Chelsea Finn and Quan Vuong and Ted Xiao},\n booktitle = {ICLR 2024},\n year = {2024},\n url = {https://proceedings.iclr.cc/paper_files/paper/2024/hash/c383e44d9a878d1982d9abb838bd5d8a-Abstract-Conference.html}\n}", "preview": "assets/papers/rt-trajectory.png", "previewSourceUrl": "https://proceedings.iclr.cc/paper_files/paper/2024/file/c383e44d9a878d1982d9abb838bd5d8a-Paper-Conference.pdf" }, "metrics": { "citations": { "value": 5, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4388512477", "updatedAt": "2026-10-05", "recordTitle": "RT-Trajectory: Robotic Task Generalization via Hindsight Trajectory Sketches" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "code-as-policies", "institutions": [ { "name": "Google Research", "url": "https://research.google/", "logo": "assets/institutions/google-research.svg", "mark": "G" } ], "name": "Code as Policies", "title": "Code as Policies: Language Model Programs for Embodied Control", "domains": [ "embodied" ], "kinds": [ "papers" ], "year": 2023, "venue": "ICRA 2023", "summary": "Prompts language models to write executable policy code that composes perception and control APIs for spatial and long-horizon robot tasks.", "tags": [ "Program synthesis", "Language planning" ], "environment": "Tabletop manipulation, mobile manipulation, and simulated control tasks.", "interface": "Natural-language instructions and API descriptions are translated into executable Python policies.", "limits": "Generated code inherits model errors and API assumptions; execution requires sandboxing, validation, and task-specific perception functions.", "status": "selected", "evidence": "Official bibliographic metadata checked; experiments not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "paper": "https://doi.org/10.1109/ICRA48891.2023.10160591", "project": "https://code-as-policies.github.io/", "pdf": "https://arxiv.org/pdf/2209.07753" }, "rank": 58, "selectionTrack": "peer-reviewed", "selectionReason": "Representative conference or journal paper within the five-year window; official publication record verified.", "paperType": "method", "bibliography": { "source": "Publisher", "sourceUrl": "https://doi.org/10.1109/ICRA48891.2023.10160591", "checkedAt": "2026-09-14", "authors": [ "Jacky Liang", "Wenlong Huang", "Fei Xia", "Peng Xu", "Karol Hausman", "Brian Ichter", "Pete Florence", "Andy Zeng" ], "date": "2023-05-29", "dateLabel": "Publisher record date", "bibtex": "@inproceedings{code_as_policies,\n title = {Code as Policies: Language Model Programs for Embodied Control},\n author = {Jacky Liang and Wenlong Huang and Fei Xia and Peng Xu and Karol Hausman and Brian Ichter and Pete Florence and Andy Zeng},\n booktitle = {ICRA 2023},\n year = {2023},\n url = {https://doi.org/10.1109/ICRA48891.2023.10160591}\n}", "preview": "assets/papers/code-as-policies.png", "previewSourceUrl": "https://arxiv.org/pdf/2209.07753" }, "metrics": { "citations": { "value": 674, "source": "OpenAlex", "sourceUrl": "https://openalex.org/W4383097638", "updatedAt": "2026-09-21", "recordTitle": "Code as Policies: Language Model Programs for Embodied Control" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-crafter-human", "name": "Crafter · open-world survival", "title": "Crafter · open-world survival", "domains": [ "game" ], "kinds": [ "demos" ], "year": 2022, "venue": "Official demonstration", "summary": "An official Crafter sequence showing resource gathering, survival, and crafting in the compact open-world benchmark.", "tags": [ "Open-world", "Game control" ], "environment": "Crafter procedural survival world.", "interface": "Pixels and discrete movement, interaction, sleep, and crafting actions.", "limits": "The clip illustrates a trajectory in a compact 2D benchmark; it is not a measure of average agent performance.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://danijar.com/project/crafter/", "video": "https://danijar.com/asset/crafter/video.mp4" }, "rank": 181, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://danijar.com/asset/crafter/video.mp4", "type": "video/mp4" }, "demoType": "game-control", "credit": "Crafter research team", "relatedIds": [ "crafter" ], "metrics": { "github": { "value": 590, "source": "GitHub", "sourceUrl": "https://github.com/danijar/crafter/stargazers", "repository": "danijar/crafter", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-vima-imitation", "name": "VIMA · one-shot video imitation", "title": "VIMA · one-shot video imitation", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2023, "venue": "Official demonstration", "summary": "An official VIMA-Bench example in which a multimodal prompt demonstrates the motion the simulated robot should reproduce.", "tags": [ "Simulation", "Multimodal prompting" ], "environment": "VIMA-Bench tabletop simulation.", "interface": "A one-shot video prompt and scene observations condition object-centric robot actions.", "limits": "The clip uses a simulator and benchmark object representation; it does not demonstrate direct transfer to a physical robot.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://vimalabs.github.io/", "video": "https://vimalabs.github.io/assets/videos/demos/one_shot_video_imitation/follow_motion/1.mp4" }, "rank": 182, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://vimalabs.github.io/assets/videos/demos/one_shot_video_imitation/follow_motion/1.mp4", "type": "video/mp4" }, "demoType": "simulation", "credit": "VIMA research team", "relatedIds": [ "vima" ], "metrics": { "github": { "value": 858, "source": "GitHub", "sourceUrl": "https://github.com/vimalabs/VIMA/stargazers", "repository": "vimalabs/VIMA", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-mobile-aloha-shrimp", "name": "Mobile ALOHA · cooking shrimp", "title": "Mobile ALOHA · cooking shrimp", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "The official Mobile ALOHA gallery shows autonomous whole-body bimanual cooking and serving of shrimp.", "tags": [ "Mobile manipulation", "Bimanual control" ], "environment": "A real kitchen with a mobile dual-arm ALOHA platform.", "interface": "Camera observations drive learned base, arm, and gripper actions cloned from whole-body demonstrations.", "limits": "A selected successful rollout does not report the task success distribution or cover changes outside the evaluated kitchen setup.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://mobile-aloha.github.io/", "video": "https://mobile-aloha.github.io/resources/skills/cook_shrimp.mp4" }, "rank": 183, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://mobile-aloha.github.io/resources/skills/cook_shrimp.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "Mobile ALOHA research team", "relatedIds": [ "mobile-aloha" ], "metrics": { "github": { "value": 4469, "source": "GitHub", "sourceUrl": "https://github.com/MarkFzp/mobile-aloha/stargazers", "repository": "MarkFzp/mobile-aloha", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-dp3-pour", "name": "DP3 · spatially robust pouring", "title": "DP3 · spatially robust pouring", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "An official real-robot DP3 sequence demonstrates point-cloud-conditioned pouring under spatial variation.", "tags": [ "3D policy", "Real robot" ], "environment": "Real tabletop pouring with sparse point-cloud observations.", "interface": "Depth-derived point clouds and proprioception condition a diffusion policy over action chunks.", "limits": "The gallery clip is qualitative evidence from the authors and does not expose the full evaluation distribution.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://3d-diffusion-policy.github.io/", "video": "https://3d-diffusion-policy.github.io/videos/high_res/pour_robust_final.mp4" }, "rank": 184, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://3d-diffusion-policy.github.io/videos/high_res/pour_robust_final.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "3D Diffusion Policy research team", "relatedIds": [ "dp3" ], "metrics": { "github": { "value": 1440, "source": "GitHub", "sourceUrl": "https://github.com/YanjieZe/3D-Diffusion-Policy/stargazers", "repository": "YanjieZe/3D-Diffusion-Policy", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-gr1-drawer", "name": "GR-1 · opening a drawer", "title": "GR-1 · opening a drawer", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "The official GR-1 page shows an end-to-end real-robot drawer-opening behavior after video-generative pretraining and policy fine-tuning.", "tags": [ "Generative policy", "Real robot" ], "environment": "Real tabletop articulated-object manipulation.", "interface": "Language, image history, and robot state tokens condition action and future-image predictions.", "limits": "This is one author-selected rollout; it does not isolate how much performance comes from video pretraining versus robot fine-tuning.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://gr1-manipulation.github.io/", "video": "https://gr1-manipulation.github.io/static/videos/real_exp/open_the_drawer.mp4" }, "rank": 185, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://gr1-manipulation.github.io/static/videos/real_exp/open_the_drawer.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "GR-1 research team", "relatedIds": [ "gr1" ], "metrics": { "github": { "value": 309, "source": "GitHub", "sourceUrl": "https://github.com/bytedance/GR-1/stargazers", "repository": "bytedance/GR-1", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-saycan-long-horizon", "name": "SayCan · long-horizon instruction", "title": "SayCan · long-horizon instruction", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2022, "venue": "Official demonstration", "summary": "An official SayCan demonstration shows language planning grounded in the feasible skills of a mobile manipulation robot.", "tags": [ "Language planning", "Affordances" ], "environment": "A real office kitchen with a mobile manipulator.", "interface": "A natural-language request is decomposed into value-weighted skills and executed sequentially.", "limits": "The planner selects from a fixed skill library; a successful sequence does not imply open-ended low-level control.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://say-can.github.io/", "video": "https://say-can.github.io/img/demo_sequence_compressed.mp4" }, "rank": 186, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://say-can.github.io/img/demo_sequence_compressed.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "SayCan research team", "relatedIds": [ "saycan" ], "metrics": {}, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-3d-diffuser-peg", "name": "3D Diffuser Actor · peg insertion", "title": "3D Diffuser Actor · peg insertion", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2024, "venue": "Official demonstration", "summary": "The official gallery shows a 3D trajectory diffusion policy inserting a peg into a hole on a real robot.", "tags": [ "3D policy", "Diffusion models" ], "environment": "Real tabletop contact-rich manipulation.", "interface": "RGB-D scene features, language, and proprioception condition a denoised end-effector trajectory.", "limits": "A gallery success does not establish tolerance across unseen geometry, calibration error, or contact conditions.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://3d-diffuser-actor.github.io/", "video": "https://3d-diffuser-actor.github.io/static/videos/realworld/new_insert_peg_in_hole_1.mp4" }, "rank": 187, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://3d-diffuser-actor.github.io/static/videos/realworld/new_insert_peg_in_hole_1.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "3D Diffuser Actor research team", "relatedIds": [ "3d-diffuser-actor" ], "metrics": {}, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-cliport-folding", "name": "CLIPort · cloth folding", "title": "CLIPort · cloth folding", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2021, "venue": "Official demonstration", "summary": "An official CLIPort real-robot clip shows language-conditioned cloth folding with semantic and spatial action pathways.", "tags": [ "Vision-language", "Real robot" ], "environment": "Real tabletop cloth manipulation.", "interface": "RGB observations and language condition pick-and-place predictions.", "limits": "The action abstraction is tailored to tabletop pick-and-place and cannot represent every deformation or continuous force profile.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://cliport.github.io/", "video": "https://cliport.github.io/media/videos/1_folding.mp4" }, "rank": 188, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://cliport.github.io/media/videos/1_folding.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "CLIPort research team", "relatedIds": [ "cliport" ], "metrics": { "github": { "value": 549, "source": "GitHub", "sourceUrl": "https://github.com/cliport/cliport/stargazers", "repository": "cliport/cliport", "updatedAt": "2026-09-14" } }, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" }, { "id": "demo-inner-monologue-recovery", "name": "Inner Monologue · feedback-driven recovery", "title": "Inner Monologue · feedback-driven recovery", "domains": [ "embodied" ], "kinds": [ "demos" ], "year": 2022, "venue": "Official demonstration", "summary": "An official example shows a language planner using scene and success feedback to update a multi-step manipulation plan.", "tags": [ "Closed-loop reasoning", "Language planning" ], "environment": "Real or simulated tabletop rearrangement from the official evaluation gallery.", "interface": "Textual environment feedback is returned to the language model between skill executions.", "limits": "Recovery depends on the accuracy of external feedback modules and the coverage of the available skill library.", "status": "selected", "evidence": "Original project video and source page checked. Qualitative, author-reported evidence; not independently reproduced.", "sourceType": "supplement", "checkedAt": "2026-09-14", "links": { "project": "https://innermonologue.github.io/", "video": "https://innermonologue.github.io/img/im_demo1_an.mp4" }, "rank": 189, "selectionTrack": "demonstration", "selectionReason": "Publicly accessible original research demonstration with an explicit setting and related peer-reviewed paper.", "video": { "src": "https://innermonologue.github.io/img/im_demo1_an.mp4", "type": "video/mp4" }, "demoType": "real-robot", "credit": "Inner Monologue research team", "relatedIds": [ "inner-monologue" ], "metrics": {}, "addedAt": "2026-09-14", "updatedAt": "2026-09-14" } ] }