{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ISWI67OZK2A6NKXLOSZVBKWFR5","short_pith_number":"pith:ISWI67OZ","schema_version":"1.0","canonical_sha256":"44ac8f7dd95681e6aaeb74b350aac58f707795e0b44b65928411941776db9282","source":{"kind":"arxiv","id":"2508.03944","version":1},"attestation_state":"computed","paper":{"title":"Constraint-Preserving Data Generation for Visuomotor Policy Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Aaditya Prasad, Andrew McAlinden, Jeannette Bohg, Jimmy Wu, Kevin Lin, Varun Ragunath, Yuke Zhu","submitted_at":"2025-08-05T22:20:02Z","abstract_excerpt":"Large-scale demonstration data has powered key breakthroughs in robot manipulation, but collecting that data remains costly and time-consuming. We present Constraint-Preserving Data Generation (CP-Gen), a method that uses a single expert trajectory to generate robot demonstrations containing novel object geometries and poses. These generated demonstrations are used to train closed-loop visuomotor policies that transfer zero-shot to the real world and generalize across variations in object geometries and poses. Similar to prior work using pose variations for data generation, CP-Gen first decomp"},"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":"2508.03944","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2025-08-05T22:20:02Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"df83f69f1e7817c67224242ab3c828bb7dc64a20e22a0602a1ddea146293367e","abstract_canon_sha256":"4292e0ed3b8e20f97a273101ffe4f0c1912fa3905f0a96f3ca2155222d24aac4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:49:17.832537Z","signature_b64":"hHYd43o2G0lOgDPcOjL/tyQoI/av1hFaY39X1sNPPsyfKqLme760tVUu3BFrsWbbj0Qz0kctci4Tc5Ak3nI5CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"44ac8f7dd95681e6aaeb74b350aac58f707795e0b44b65928411941776db9282","last_reissued_at":"2026-07-05T11:49:17.832027Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:49:17.832027Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Constraint-Preserving Data Generation for Visuomotor Policy Learning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Aaditya Prasad, Andrew McAlinden, Jeannette Bohg, Jimmy Wu, Kevin Lin, Varun Ragunath, Yuke Zhu","submitted_at":"2025-08-05T22:20:02Z","abstract_excerpt":"Large-scale demonstration data has powered key breakthroughs in robot manipulation, but collecting that data remains costly and time-consuming. We present Constraint-Preserving Data Generation (CP-Gen), a method that uses a single expert trajectory to generate robot demonstrations containing novel object geometries and poses. These generated demonstrations are used to train closed-loop visuomotor policies that transfer zero-shot to the real world and generalize across variations in object geometries and poses. Similar to prior work using pose variations for data generation, CP-Gen first decomp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.03944","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/2508.03944/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":"2508.03944","created_at":"2026-07-05T11:49:17.832096+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.03944v1","created_at":"2026-07-05T11:49:17.832096+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.03944","created_at":"2026-07-05T11:49:17.832096+00:00"},{"alias_kind":"pith_short_12","alias_value":"ISWI67OZK2A6","created_at":"2026-07-05T11:49:17.832096+00:00"},{"alias_kind":"pith_short_16","alias_value":"ISWI67OZK2A6NKXL","created_at":"2026-07-05T11:49:17.832096+00:00"},{"alias_kind":"pith_short_8","alias_value":"ISWI67OZ","created_at":"2026-07-05T11:49:17.832096+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.13645","citing_title":"A Mechanistic Analysis of Sim-and-Real Co-Training in Generative Robot Policies","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15569","citing_title":"ShapeGen: Robotic Data Generation for Category-Level Manipulation","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15023","citing_title":"DockAnywhere: Data-Efficient Visuomotor Policy Learning for Mobile Manipulation via Novel Demonstration Generation","ref_index":32,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5","json":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5.json","graph_json":"https://pith.science/api/pith-number/ISWI67OZK2A6NKXLOSZVBKWFR5/graph.json","events_json":"https://pith.science/api/pith-number/ISWI67OZK2A6NKXLOSZVBKWFR5/events.json","paper":"https://pith.science/paper/ISWI67OZ"},"agent_actions":{"view_html":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5","download_json":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5.json","view_paper":"https://pith.science/paper/ISWI67OZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.03944&json=true","fetch_graph":"https://pith.science/api/pith-number/ISWI67OZK2A6NKXLOSZVBKWFR5/graph.json","fetch_events":"https://pith.science/api/pith-number/ISWI67OZK2A6NKXLOSZVBKWFR5/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5/action/storage_attestation","attest_author":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5/action/author_attestation","sign_citation":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5/action/citation_signature","submit_replication":"https://pith.science/pith/ISWI67OZK2A6NKXLOSZVBKWFR5/action/replication_record"}},"created_at":"2026-07-05T11:49:17.832096+00:00","updated_at":"2026-07-05T11:49:17.832096+00:00"}