{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4IODRXPS6XDVUY4JCCFIB7LMBP","short_pith_number":"pith:4IODRXPS","schema_version":"1.0","canonical_sha256":"e21c38ddf2f5c75a6389108a80fd6c0bc071ca13aa4fc90398dd1c3c7cab9103","source":{"kind":"arxiv","id":"2409.08273","version":1},"attestation_state":"computed","paper":{"title":"Hand-Object Interaction Pretraining from Videos","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.RO","authors_text":"Antonio Loquercio, Carmelo Sferrazza, Haozhi Qi, Himanshu Gaurav Singh, Jane Wu, Jitendra Malik, Pieter Abbeel","submitted_at":"2024-09-12T17:59:07Z","abstract_excerpt":"We present an approach to learn general robot manipulation priors from 3D hand-object interaction trajectories. We build a framework to use in-the-wild videos to generate sensorimotor robot trajectories. We do so by lifting both the human hand and the manipulated object in a shared 3D space and retargeting human motions to robot actions. Generative modeling on this data gives us a task-agnostic base policy. This policy captures a general yet flexible manipulation prior. We empirically demonstrate that finetuning this policy, with both reinforcement learning (RL) and behavior cloning (BC), enab"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2409.08273","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.RO","submitted_at":"2024-09-12T17:59:07Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"14add8561aaad2379ce587ccaf053f814a2e23b8ddf5f77becafdd13d6bb85eb","abstract_canon_sha256":"4c27b765cc76b451f82fc7660d5f2ada0c6d2ccb018def6339f6750736dde6ca"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:06:22.657975Z","signature_b64":"QU7iiofdkIRW7VJbw4Pen8QbPg3h+xuGrWvMs68D+2v4Y4+E+CgenarzaRigGd/rRvOTs8XNkn+rZw3S726ZDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e21c38ddf2f5c75a6389108a80fd6c0bc071ca13aa4fc90398dd1c3c7cab9103","last_reissued_at":"2026-07-05T09:06:22.657521Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:06:22.657521Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Hand-Object Interaction Pretraining from Videos","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.RO","authors_text":"Antonio Loquercio, Carmelo Sferrazza, Haozhi Qi, Himanshu Gaurav Singh, Jane Wu, Jitendra Malik, Pieter Abbeel","submitted_at":"2024-09-12T17:59:07Z","abstract_excerpt":"We present an approach to learn general robot manipulation priors from 3D hand-object interaction trajectories. We build a framework to use in-the-wild videos to generate sensorimotor robot trajectories. We do so by lifting both the human hand and the manipulated object in a shared 3D space and retargeting human motions to robot actions. Generative modeling on this data gives us a task-agnostic base policy. This policy captures a general yet flexible manipulation prior. We empirically demonstrate that finetuning this policy, with both reinforcement learning (RL) and behavior cloning (BC), enab"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.08273","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2409.08273/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2409.08273","created_at":"2026-07-05T09:06:22.657579+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.08273v1","created_at":"2026-07-05T09:06:22.657579+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.08273","created_at":"2026-07-05T09:06:22.657579+00:00"},{"alias_kind":"pith_short_12","alias_value":"4IODRXPS6XDV","created_at":"2026-07-05T09:06:22.657579+00:00"},{"alias_kind":"pith_short_16","alias_value":"4IODRXPS6XDVUY4J","created_at":"2026-07-05T09:06:22.657579+00:00"},{"alias_kind":"pith_short_8","alias_value":"4IODRXPS","created_at":"2026-07-05T09:06:22.657579+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19333","citing_title":"Do as I Do: Dexterous Manipulation Data from Everyday Human Videos","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2505.07813","citing_title":"DexWild: Dexterous Human Interactions for In-the-Wild Robot Policies","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2501.09747","citing_title":"FAST: Efficient Action Tokenization for Vision-Language-Action Models","ref_index":58,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP","json":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP.json","graph_json":"https://pith.science/api/pith-number/4IODRXPS6XDVUY4JCCFIB7LMBP/graph.json","events_json":"https://pith.science/api/pith-number/4IODRXPS6XDVUY4JCCFIB7LMBP/events.json","paper":"https://pith.science/paper/4IODRXPS"},"agent_actions":{"view_html":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP","download_json":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP.json","view_paper":"https://pith.science/paper/4IODRXPS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.08273&json=true","fetch_graph":"https://pith.science/api/pith-number/4IODRXPS6XDVUY4JCCFIB7LMBP/graph.json","fetch_events":"https://pith.science/api/pith-number/4IODRXPS6XDVUY4JCCFIB7LMBP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP/action/storage_attestation","attest_author":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP/action/author_attestation","sign_citation":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP/action/citation_signature","submit_replication":"https://pith.science/pith/4IODRXPS6XDVUY4JCCFIB7LMBP/action/replication_record"}},"created_at":"2026-07-05T09:06:22.657579+00:00","updated_at":"2026-07-05T09:06:22.657579+00:00"}