{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:K4JLMF4VGLVUSGN3F5MWXCDG6P","short_pith_number":"pith:K4JLMF4V","schema_version":"1.0","canonical_sha256":"5712b6179532eb4919bb2f596b8866f3d163aebeb5f528b703d27a71f90118b2","source":{"kind":"arxiv","id":"2111.03062","version":1},"attestation_state":"computed","paper":{"title":"Generalization in Dexterous Manipulation via Geometry-Aware Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.LG","cs.SY","eess.SY"],"primary_cat":"cs.RO","authors_text":"Deepak Pathak, Igor Mordatch, Pieter Abbeel, Wenlong Huang","submitted_at":"2021-11-04T17:59:56Z","abstract_excerpt":"Dexterous manipulation of arbitrary objects, a fundamental daily task for humans, has been a grand challenge for autonomous robotic systems. Although data-driven approaches using reinforcement learning can develop specialist policies that discover behaviors to control a single object, they often exhibit poor generalization to unseen ones. In this work, we show that policies learned by existing reinforcement learning algorithms can in fact be generalist when combined with multi-task learning and a well-chosen object representation. We show that a single generalist policy can perform in-hand man"},"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":"2111.03062","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2021-11-04T17:59:56Z","cross_cats_sorted":["cs.AI","cs.CV","cs.LG","cs.SY","eess.SY"],"title_canon_sha256":"92ad533eb829b9a90cbeb7e4dddd7cf60e8cf1a78b6a755c13d80a41435f9bb6","abstract_canon_sha256":"4d79f8171068ebf496594ee3b30e42da451cbd532779ed40034884f9376dc32e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:29:12.007331Z","signature_b64":"+WPS9r4P+tFYf9j9YHxk3a+rKHLATmb6nWm0b4Uv6vzvDEZ4yvsLbqjqp4T87QSr6Ttx8m52nlFJBY52QDpHDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5712b6179532eb4919bb2f596b8866f3d163aebeb5f528b703d27a71f90118b2","last_reissued_at":"2026-07-05T03:29:12.006856Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:29:12.006856Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generalization in Dexterous Manipulation via Geometry-Aware Multi-Task Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.LG","cs.SY","eess.SY"],"primary_cat":"cs.RO","authors_text":"Deepak Pathak, Igor Mordatch, Pieter Abbeel, Wenlong Huang","submitted_at":"2021-11-04T17:59:56Z","abstract_excerpt":"Dexterous manipulation of arbitrary objects, a fundamental daily task for humans, has been a grand challenge for autonomous robotic systems. Although data-driven approaches using reinforcement learning can develop specialist policies that discover behaviors to control a single object, they often exhibit poor generalization to unseen ones. In this work, we show that policies learned by existing reinforcement learning algorithms can in fact be generalist when combined with multi-task learning and a well-chosen object representation. We show that a single generalist policy can perform in-hand man"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2111.03062","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/2111.03062/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":"2111.03062","created_at":"2026-07-05T03:29:12.006913+00:00"},{"alias_kind":"arxiv_version","alias_value":"2111.03062v1","created_at":"2026-07-05T03:29:12.006913+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2111.03062","created_at":"2026-07-05T03:29:12.006913+00:00"},{"alias_kind":"pith_short_12","alias_value":"K4JLMF4VGLVU","created_at":"2026-07-05T03:29:12.006913+00:00"},{"alias_kind":"pith_short_16","alias_value":"K4JLMF4VGLVUSGN3","created_at":"2026-07-05T03:29:12.006913+00:00"},{"alias_kind":"pith_short_8","alias_value":"K4JLMF4V","created_at":"2026-07-05T03:29:12.006913+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21788","citing_title":"Rotation-Aware Point-Cloud Embeddings for Vision-Based In-Hand Reorientation","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12759","citing_title":"Sparse2Act: Learning Action-Aligned Sparse 3D Representations for Cross-Domain Robot Manipulation","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03268","citing_title":"EaDex: A Cross-Embodiment Dexterous Manipulation Framework from Low-Cost Demonstrations","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17017","citing_title":"When Dynamics Shift, Robust Task Inference Wins: Offline Imitation Learning with Behavior Foundation Models Revisited","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09157","citing_title":"Revisiting Mixture Policies in Entropy-Regularized Actor-Critic","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P","json":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P.json","graph_json":"https://pith.science/api/pith-number/K4JLMF4VGLVUSGN3F5MWXCDG6P/graph.json","events_json":"https://pith.science/api/pith-number/K4JLMF4VGLVUSGN3F5MWXCDG6P/events.json","paper":"https://pith.science/paper/K4JLMF4V"},"agent_actions":{"view_html":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P","download_json":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P.json","view_paper":"https://pith.science/paper/K4JLMF4V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2111.03062&json=true","fetch_graph":"https://pith.science/api/pith-number/K4JLMF4VGLVUSGN3F5MWXCDG6P/graph.json","fetch_events":"https://pith.science/api/pith-number/K4JLMF4VGLVUSGN3F5MWXCDG6P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P/action/storage_attestation","attest_author":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P/action/author_attestation","sign_citation":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P/action/citation_signature","submit_replication":"https://pith.science/pith/K4JLMF4VGLVUSGN3F5MWXCDG6P/action/replication_record"}},"created_at":"2026-07-05T03:29:12.006913+00:00","updated_at":"2026-07-05T03:29:12.006913+00:00"}