{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZVET4CO4HJ5WYAFV72WKEXH6J7","short_pith_number":"pith:ZVET4CO4","schema_version":"1.0","canonical_sha256":"cd493e09dc3a7b6c00b5feaca25cfe4fda55533c0ad8dbe2edc20a23ef779d1c","source":{"kind":"arxiv","id":"2412.06685","version":1},"attestation_state":"computed","paper":{"title":"Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Archit Sharma, Aviral Kumar, Chelsea Finn, Georgia Gabriela Sampaio, Max Sobol Mark, Mohan Kumar Srirama, Tian Gao","submitted_at":"2024-12-09T17:28:03Z","abstract_excerpt":"Recent advances in learning decision-making policies can largely be attributed to training expressive policy models, largely via imitation learning. While imitation learning discards non-expert data, reinforcement learning (RL) can still learn from suboptimal data. However, instantiating RL training of a new policy class often presents a different challenge: most deep RL machinery is co-developed with assumptions on the policy class and backbone, resulting in poor performance when the policy class changes. For instance, SAC utilizes a low-variance reparameterization policy gradient for Gaussia"},"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":"2412.06685","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-12-09T17:28:03Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8cd84b0f65921cfda78015bf1a2b701e47e3d99372d20c6dfc7429f327782f4f","abstract_canon_sha256":"3fe3e99c7b4976418135c469e55ce58c2a6319ac1c2de8d6aecdb51c45e2d755"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:46:36.475403Z","signature_b64":"QvxvwOhJo3/swM7icNHbMMLYrMlLpjP7e0GdRkWuHVT3V4F0CnLaps8JItairYuW2/JzLEh/YO9lyX/QgU7oAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cd493e09dc3a7b6c00b5feaca25cfe4fda55533c0ad8dbe2edc20a23ef779d1c","last_reissued_at":"2026-07-05T09:46:36.474916Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:46:36.474916Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Policy Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Archit Sharma, Aviral Kumar, Chelsea Finn, Georgia Gabriela Sampaio, Max Sobol Mark, Mohan Kumar Srirama, Tian Gao","submitted_at":"2024-12-09T17:28:03Z","abstract_excerpt":"Recent advances in learning decision-making policies can largely be attributed to training expressive policy models, largely via imitation learning. While imitation learning discards non-expert data, reinforcement learning (RL) can still learn from suboptimal data. However, instantiating RL training of a new policy class often presents a different challenge: most deep RL machinery is co-developed with assumptions on the policy class and backbone, resulting in poor performance when the policy class changes. For instance, SAC utilizes a low-variance reparameterization policy gradient for Gaussia"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.06685","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/2412.06685/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":"2412.06685","created_at":"2026-07-05T09:46:36.474975+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.06685v1","created_at":"2026-07-05T09:46:36.474975+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.06685","created_at":"2026-07-05T09:46:36.474975+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZVET4CO4HJ5W","created_at":"2026-07-05T09:46:36.474975+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZVET4CO4HJ5WYAFV","created_at":"2026-07-05T09:46:36.474975+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZVET4CO4","created_at":"2026-07-05T09:46:36.474975+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":21,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23640","citing_title":"Learning Process Rewards via Success Visitation Matching for Efficient RL","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2606.17551","citing_title":"Reversal Q-Learning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13675","citing_title":"Improving Robotic Generalist Policies via Flow Reversal Steering","ref_index":49,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13795","citing_title":"DiPOD: Diffusion Policy Optimization without Drifting Apart","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.10825","citing_title":"MODIP: Efficient Model-Based Optimization for Diffusion Policies","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11087","citing_title":"Test-Time Gradient Guidance of Flow Policies in Reinforcement Learning","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31958","citing_title":"Adapting Generalist Robot Policies with Semantic Reinforcement Learning","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03065","citing_title":"OGPO: Sample Efficient Full-Finetuning of Generative Control Policies","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25477","citing_title":"EXPO-FT: Sample-Efficient Reinforcement Learning Finetuning for Vision-Language-Action Models","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2602.07399","citing_title":"VGAS: Value-Guided Action-Chunk Selection for Few-Shot Vision-Language-Action Adaptation","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22376","citing_title":"Target-Aligned Bellman Backup for Cross-domain Offline Reinforcement Learning","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2603.12243","citing_title":"HandelBot: Real-World Piano Playing via Fast Adaptation of Dexterous Robot Policies","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2506.15799","citing_title":"Steering Your Diffusion Policy with Latent Space Reinforcement Learning","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2603.12243","citing_title":"HandelBot: Real-World Piano Playing via Fast Adaptation of Dexterous Robot Policies","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07637","citing_title":"Learning to Communicate Locally for Large-Scale Multi-Agent Pathfinding","ref_index":80,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11832","citing_title":"Learning Action Manifold with Multi-view Latent Priors for Robotic Manipulation","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2511.14759","citing_title":"$\\pi^{*}_{0.6}$: a VLA That Learns From Experience","ref_index":39,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07637","citing_title":"Learning to Communicate Locally for Large-Scale Multi-Agent Pathfinding","ref_index":80,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07381","citing_title":"Escaping the Diversity Trap in Robotic Manipulation via Anchor-Centric Adaptation","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14265","citing_title":"Reinforcement Learning via Value Gradient Flow","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03065","citing_title":"OGPO: Sample Efficient Full-Finetuning of Generative Control Policies","ref_index":160,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7","json":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7.json","graph_json":"https://pith.science/api/pith-number/ZVET4CO4HJ5WYAFV72WKEXH6J7/graph.json","events_json":"https://pith.science/api/pith-number/ZVET4CO4HJ5WYAFV72WKEXH6J7/events.json","paper":"https://pith.science/paper/ZVET4CO4"},"agent_actions":{"view_html":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7","download_json":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7.json","view_paper":"https://pith.science/paper/ZVET4CO4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.06685&json=true","fetch_graph":"https://pith.science/api/pith-number/ZVET4CO4HJ5WYAFV72WKEXH6J7/graph.json","fetch_events":"https://pith.science/api/pith-number/ZVET4CO4HJ5WYAFV72WKEXH6J7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7/action/storage_attestation","attest_author":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7/action/author_attestation","sign_citation":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7/action/citation_signature","submit_replication":"https://pith.science/pith/ZVET4CO4HJ5WYAFV72WKEXH6J7/action/replication_record"}},"created_at":"2026-07-05T09:46:36.474975+00:00","updated_at":"2026-07-05T09:46:36.474975+00:00"}