{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ZO3TO2T6MWNCEK2QKZBJVE34QJ","short_pith_number":"pith:ZO3TO2T6","schema_version":"1.0","canonical_sha256":"cbb7376a7e659a222b5056429a937c826a2227b3a80b979ee937cf1e74f49439","source":{"kind":"arxiv","id":"2603.10263","version":2},"attestation_state":"computed","paper":{"title":"From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.RO","authors_text":"Shuran Song, Zhanyi Sun","submitted_at":"2026-03-10T22:49:46Z","abstract_excerpt":"We introduce Distribution Contractive Reinforcement Learning (DICE-RL), a framework that uses reinforcement learning (RL) as a \"distribution contraction\" operator to refine pretrained generative robot policies. DICE-RL turns a pretrained behavior prior into a high-performing \"pro\" policy by amplifying high-success behaviors from online feedback. We pretrain a diffusion- or flow-based policy for broad behavioral coverage, then finetune it with a stable, sample-efficient residual off-policy RL framework that combines selective behavior regularization with value-guided action selection. Extensive"},"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":"2603.10263","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2026-03-10T22:49:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"58a54e9944d049fee997e9cd07e806fa100292f562d0cf399f6e38ecc17e4a27","abstract_canon_sha256":"b185847398d8a5b1323169be882bf9528cc2ec549493b4061c4ca8aca9b96f3c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-02T01:17:29.309462Z","signature_b64":"AUwmITRUuw2Nw2o5r0MOfSzNMjWaBfRHFwGZGh7kxOAdmOt7h4yLHaPJqNHUmz6i3Feu7KbeNF9q8zy1jzgDBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cbb7376a7e659a222b5056429a937c826a2227b3a80b979ee937cf1e74f49439","last_reissued_at":"2026-07-02T01:17:29.308942Z","signature_status":"signed_v1","first_computed_at":"2026-07-02T01:17:29.308942Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Prior to Pro: Efficient Skill Mastery via Distribution Contractive RL Finetuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.RO","authors_text":"Shuran Song, Zhanyi Sun","submitted_at":"2026-03-10T22:49:46Z","abstract_excerpt":"We introduce Distribution Contractive Reinforcement Learning (DICE-RL), a framework that uses reinforcement learning (RL) as a \"distribution contraction\" operator to refine pretrained generative robot policies. DICE-RL turns a pretrained behavior prior into a high-performing \"pro\" policy by amplifying high-success behaviors from online feedback. We pretrain a diffusion- or flow-based policy for broad behavioral coverage, then finetune it with a stable, sample-efficient residual off-policy RL framework that combines selective behavior regularization with value-guided action selection. Extensive"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.10263","kind":"arxiv","version":2},"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/2603.10263/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":"2603.10263","created_at":"2026-07-02T01:17:29.309000+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.10263v2","created_at":"2026-07-02T01:17:29.309000+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.10263","created_at":"2026-07-02T01:17:29.309000+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZO3TO2T6MWNC","created_at":"2026-07-02T01:17:29.309000+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZO3TO2T6MWNCEK2Q","created_at":"2026-07-02T01:17:29.309000+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZO3TO2T6","created_at":"2026-07-02T01:17:29.309000+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2606.02194","citing_title":"Coherent Off-Policy Improvement of Large Behavior Models with Learned Rewards","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2606.31958","citing_title":"Adapting Generalist Robot Policies with Semantic Reinforcement Learning","ref_index":19,"is_internal_anchor":true},{"citing_arxiv_id":"2605.19919","citing_title":"Beyond Action Residuals: Real-World Robot Policy Steering via Bottleneck Latent Reinforcement Learning","ref_index":51,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ","json":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ.json","graph_json":"https://pith.science/api/pith-number/ZO3TO2T6MWNCEK2QKZBJVE34QJ/graph.json","events_json":"https://pith.science/api/pith-number/ZO3TO2T6MWNCEK2QKZBJVE34QJ/events.json","paper":"https://pith.science/paper/ZO3TO2T6"},"agent_actions":{"view_html":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ","download_json":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ.json","view_paper":"https://pith.science/paper/ZO3TO2T6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.10263&json=true","fetch_graph":"https://pith.science/api/pith-number/ZO3TO2T6MWNCEK2QKZBJVE34QJ/graph.json","fetch_events":"https://pith.science/api/pith-number/ZO3TO2T6MWNCEK2QKZBJVE34QJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ/action/storage_attestation","attest_author":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ/action/author_attestation","sign_citation":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ/action/citation_signature","submit_replication":"https://pith.science/pith/ZO3TO2T6MWNCEK2QKZBJVE34QJ/action/replication_record"}},"created_at":"2026-07-02T01:17:29.309000+00:00","updated_at":"2026-07-02T01:17:29.309000+00:00"}