Librarian Bot. I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API

FoundationMotion: Auto-Labeling and Reasoning about Spatial Movement in Videos (2025)
Zero-Shot Open-Vocabulary Human Motion Grounding with Test-Time Training (2025)
R-AVST: Empowering Video-LLMs with Fine-Grained Spatio-Temporal Reasoning in Complex Audio-Visual Scenarios (2025)
SOVABench: A Vehicle Surveillance Action Retrieval Benchmark for Multimodal Large Language Models (2026)
InternVideo-Next: Towards General Video Foundation Models without Video-Text Supervision (2025)
OneStory: Coherent Multi-Shot Video Generation with Adaptive Memory (2025)
LongVideoAgent: Multi-Agent Reasoning with Long Videos (2025)

Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space

You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: \n\n@librarian-bot\n\t recommend

\n","updatedAt":"2026-01-17T01:37:47.506Z","author":{"_id":"63d3e0e8ff1384ce6c5dd17d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg","fullname":"Librarian Bot (Bot)","name":"librarian-bot","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":318,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6909293532371521},"editors":["librarian-bot"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/1674830754237-63d3e0e8ff1384ce6c5dd17d.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2601.10592","authors":[{"_id":"6969b18332f0333869ff9484","user":{"_id":"630491107424d937fa3258be","avatarUrl":"/avatars/b8bd81bc8544674ee26b78702afdb87c.svg","isPro":true,"fullname":"Delong Chen","user":"chendelong","type":"user"},"name":"Delong Chen","status":"claimed_verified","statusLastChangedAt":"2026-01-16T14:37:03.915Z","hidden":false},{"_id":"6969b18332f0333869ff9485","user":{"_id":"67a4de92a41dcedc1632eb43","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/rKpXABi8RqSeRh2RXBlZU.png","isPro":false,"fullname":"Tejaswi Kasarla","user":"tkasarla","type":"user"},"name":"Tejaswi Kasarla","status":"admin_assigned","statusLastChangedAt":"2026-01-16T15:52:21.274Z","hidden":false},{"_id":"6969b18332f0333869ff9486","user":{"_id":"6566fd7447b037bd21f4147e","avatarUrl":"/avatars/c0da0ace33c6c163ec5fa5a866e1296b.svg","isPro":false,"fullname":"Yejin Bang","user":"yjbang","type":"user"},"name":"Yejin Bang","status":"claimed_verified","statusLastChangedAt":"2026-01-21T09:25:09.737Z","hidden":false},{"_id":"6969b18332f0333869ff9487","user":{"_id":"62bdeedd01dc22b4d22a371e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/62bdeedd01dc22b4d22a371e/ahbK9Ehurx1TgQAVw1TcS.jpeg","isPro":false,"fullname":"Mustafa Shukor","user":"mshukor","type":"user"},"name":"Mustafa Shukor","status":"admin_assigned","statusLastChangedAt":"2026-01-16T15:52:30.924Z","hidden":false},{"_id":"6969b18332f0333869ff9488","user":{"_id":"681abde923dbf6b59c410214","avatarUrl":"/avatars/f863d0a3d05fc99b29535d797d7a6292.svg","isPro":false,"fullname":"willy Chung","user":"willy1027","type":"user"},"name":"Willy Chung","status":"admin_assigned","statusLastChangedAt":"2026-01-16T15:52:37.777Z","hidden":false},{"_id":"6969b18332f0333869ff9489","name":"Jade Yu","hidden":false},{"_id":"6969b18332f0333869ff948a","user":{"_id":"65f08f0bcd30f1ed6ea6cbd1","avatarUrl":"/avatars/5eb2c2032dc2054d4ca86665ba0d428c.svg","isPro":false,"fullname":"Allen Bolourchi","user":"allen-ml","type":"user"},"name":"Allen Bolourchi","status":"admin_assigned","statusLastChangedAt":"2026-01-16T15:52:48.723Z","hidden":false},{"_id":"6969b18332f0333869ff948b","user":{"_id":"639111acb73a0f8c02a92844","avatarUrl":"/avatars/58766701471be66fd784bedd319741a4.svg","isPro":false,"fullname":"Theo Moutakanni","user":"TheoM","type":"user"},"name":"Theo Moutakanni","status":"admin_assigned","statusLastChangedAt":"2026-01-16T15:52:55.020Z","hidden":false},{"_id":"6969b18332f0333869ff948c","name":"Pascale Fung","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/6039478ab3ecf716b1a5fd4d/G3oU0qIaKddVQVgjhv5A2.gif"],"publishedAt":"2026-01-15T17:02:27.000Z","submittedOnDailyAt":"2026-01-16T01:03:48.816Z","title":"Action100M: A Large-scale Video Action Dataset","submittedOnDailyBy":{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},"summary":"Inferring physical actions from visual observations is a fundamental capability for advancing machine intelligence in the physical world. Achieving this requires large-scale, open-vocabulary video action datasets that span broad domains. We introduce Action100M, a large-scale dataset constructed from 1.2M Internet instructional videos (14.6 years of duration), yielding O(100 million) temporally localized segments with open-vocabulary action supervision and rich captions. Action100M is generated by a fully automated pipeline that (i) performs hierarchical temporal segmentation using V-JEPA 2 embeddings, (ii) produces multi-level frame and segment captions organized as a Tree-of-Captions, and (iii) aggregates evidence with a reasoning model (GPT-OSS-120B) under a multi-round Self-Refine procedure to output structured annotations (brief/detailed action, actor, brief/detailed caption). Training VL-JEPA on Action100M demonstrates consistent data-scaling improvements and strong zero-shot performance across diverse action recognition benchmarks, establishing Action100M as a new foundation for scalable research in video understanding and world modeling.","upvotes":29,"discussionId":"6969b18332f0333869ff948d","githubRepo":"https://github.com/facebookresearch/Action100M","githubRepoAddedBy":"user","ai_summary":"Action100M is a large-scale video action dataset constructed from internet instructional videos using automated pipelines with V-JEPA embeddings and GPT-based reasoning for structured annotations.","ai_keywords":["Action100M","V-JEPA","Tree-of-Captions","GPT-OSS-120B","Self-Refine","video action recognition","zero-shot performance","data-scaling improvements"],"githubStars":404,"organization":{"_id":"66b54027408752ae16404b05","name":"metaresearch","fullname":"Meta Research","avatar":"https://cdn-uploads.huggingface.co/production/uploads/66b25f3f58babfaeb76112dc/2GmiaF075AZ7BcE538oPk.png"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},{"_id":"63e202f352b7578dba448ab5","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e202f352b7578dba448ab5/8itVBLcv14m7OVsoF8h1o.jpeg","isPro":false,"fullname":"Kaicheng Yang","user":"Kaichengalex","type":"user"},{"_id":"652ce0d4c543a08aa92e010f","avatarUrl":"/avatars/7978304e3fe99b0d4d0712441c6a24f3.svg","isPro":false,"fullname":"Haoyu Guo","user":"ghy0324","type":"user"},{"_id":"6538764e9c474315d73c9426","avatarUrl":"/avatars/cc959caef6e4f87931e078ba90f6e423.svg","isPro":false,"fullname":"Tran Tuan Huy","user":"HuyTT","type":"user"},{"_id":"671618d2c6f6570d4f513004","avatarUrl":"/avatars/6ac6577d110731a1e8e3d56b9018fc86.svg","isPro":false,"fullname":"LiangYao","user":"1e12Leon","type":"user"},{"_id":"68e89fc2098a95a3b0a6369b","avatarUrl":"/avatars/4f5e94225ffac574d73df0495154b377.svg","isPro":false,"fullname":"Qi Wong","user":"wong-qi-1234","type":"user"},{"_id":"670a5c886f31d354bc8c1cd1","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/670a5c886f31d354bc8c1cd1/D2Mueg9eQy4fzJhOmCrnu.jpeg","isPro":false,"fullname":"Wenzheng Zeng","user":"wenzhengzeng","type":"user"},{"_id":"624ac233c04d55ec0f42b11e","avatarUrl":"/avatars/58a9abce945e71a65abc8a54085de6d7.svg","isPro":false,"fullname":"oh sehun","user":"sehun","type":"user"},{"_id":"6270324ebecab9e2dcf245de","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6270324ebecab9e2dcf245de/cMbtWSasyNlYc9hvsEEzt.jpeg","isPro":false,"fullname":"Kye Gomez","user":"kye","type":"user"},{"_id":"696a97eb54453892ee130070","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/gaHwMDNPcm-BSuKgD4PLr.png","isPro":false,"fullname":"Martha Deeks","user":"MarthaDeeks","type":"user"},{"_id":"66249e5c5e23c3941aa6cb09","avatarUrl":"/avatars/d3f68c2eba04f2698402cc532779a549.svg","isPro":false,"fullname":"Chunjiang Ge","user":"ChunjiangGe","type":"user"},{"_id":"6570450a78d7aca0c361a177","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6570450a78d7aca0c361a177/MX7jHhTQwLs-BvYIu5rqb.jpeg","isPro":false,"fullname":"Harold Chen","user":"Harold328","type":"user"}],"acceptLanguages":["*"],"dailyPaperRank":0,"organization":{"_id":"66b54027408752ae16404b05","name":"metaresearch","fullname":"Meta Research","avatar":"https://cdn-uploads.huggingface.co/production/uploads/66b25f3f58babfaeb76112dc/2GmiaF075AZ7BcE538oPk.png"}}">

Papers

arxiv:2601.10592

Action100M: A Large-scale Video Action Dataset

Published on Jan 15

· Submitted by

taesiri on Jan 16

Meta Research

Upvote

Authors:

Delong Chen ,

Tejaswi Kasarla ,

Yejin Bang ,

Mustafa Shukor ,

Willy Chung ,

Allen Bolourchi ,

Theo Moutakanni ,

Abstract

Action100M is a large-scale video action dataset constructed from internet instructional videos using automated pipelines with V-JEPA embeddings and GPT-based reasoning for structured annotations.

AI-generated summary

Inferring physical actions from visual observations is a fundamental capability for advancing machine intelligence in the physical world. Achieving this requires large-scale, open-vocabulary video action datasets that span broad domains. We introduce Action100M, a large-scale dataset constructed from 1.2M Internet instructional videos (14.6 years of duration), yielding O(100 million) temporally localized segments with open-vocabulary action supervision and rich captions. Action100M is generated by a fully automated pipeline that (i) performs hierarchical temporal segmentation using V-JEPA 2 embeddings, (ii) produces multi-level frame and segment captions organized as a Tree-of-Captions, and (iii) aggregates evidence with a reasoning model (GPT-OSS-120B) under a multi-round Self-Refine procedure to output structured annotations (brief/detailed action, actor, brief/detailed caption). Training VL-JEPA on Action100M demonstrates consistent data-scaling improvements and strong zero-shot performance across diverse action recognition benchmarks, establishing Action100M as a new foundation for scalable research in video understanding and world modeling.