{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:4WM2H6SE56AGRA65DSR655ICVB","short_pith_number":"pith:4WM2H6SE","schema_version":"1.0","canonical_sha256":"e599a3fa44ef806883dd1ca3eef502a85ceb2020a28a58471022183adf27e95d","source":{"kind":"arxiv","id":"2412.13662","version":1},"attestation_state":"computed","paper":{"title":"When Should We Prefer State-to-Visual DAgger Over Visual Reinforcement Learning?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.RO"],"primary_cat":"cs.CV","authors_text":"Hao Su, Litian Liang, Stanis{\\l}aw Wiktor Strzelecki, Tongzhou Mu, Xiu Yuan, Yunchao Yao, Zhaoyang Li","submitted_at":"2024-12-18T09:39:12Z","abstract_excerpt":"Learning policies from high-dimensional visual inputs, such as pixels and point clouds, is crucial in various applications. Visual reinforcement learning is a promising approach that directly trains policies from visual observations, although it faces challenges in sample efficiency and computational costs. This study conducts an empirical comparison of State-to-Visual DAgger, a two-stage framework that initially trains a state policy before adopting online imitation to learn a visual policy, and Visual RL across a diverse set of tasks. We evaluate both methods across 16 tasks from three bench"},"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":"2412.13662","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-18T09:39:12Z","cross_cats_sorted":["cs.AI","cs.LG","cs.RO"],"title_canon_sha256":"420db2412d520e7eabdd046cff50e7ed45b876601d5cf6993c08fddb5e45e127","abstract_canon_sha256":"43569509a01fc33ee9f837d6a15e9923c458c3f4461246e45261ca1781bc03a6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:51:06.093549Z","signature_b64":"LEXHdBivY4IAugYven3k6KspWhuU2yT0HgPQNuSI8ZCm5z+F3zbe/TL2ycBuiM8eDlneY7tghFccjnnUajlCCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e599a3fa44ef806883dd1ca3eef502a85ceb2020a28a58471022183adf27e95d","last_reissued_at":"2026-07-05T09:51:06.093078Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:51:06.093078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"When Should We Prefer State-to-Visual DAgger Over Visual Reinforcement Learning?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","cs.RO"],"primary_cat":"cs.CV","authors_text":"Hao Su, Litian Liang, Stanis{\\l}aw Wiktor Strzelecki, Tongzhou Mu, Xiu Yuan, Yunchao Yao, Zhaoyang Li","submitted_at":"2024-12-18T09:39:12Z","abstract_excerpt":"Learning policies from high-dimensional visual inputs, such as pixels and point clouds, is crucial in various applications. Visual reinforcement learning is a promising approach that directly trains policies from visual observations, although it faces challenges in sample efficiency and computational costs. This study conducts an empirical comparison of State-to-Visual DAgger, a two-stage framework that initially trains a state policy before adopting online imitation to learn a visual policy, and Visual RL across a diverse set of tasks. We evaluate both methods across 16 tasks from three bench"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.13662","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/2412.13662/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":"2412.13662","created_at":"2026-07-05T09:51:06.093134+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.13662v1","created_at":"2026-07-05T09:51:06.093134+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.13662","created_at":"2026-07-05T09:51:06.093134+00:00"},{"alias_kind":"pith_short_12","alias_value":"4WM2H6SE56AG","created_at":"2026-07-05T09:51:06.093134+00:00"},{"alias_kind":"pith_short_16","alias_value":"4WM2H6SE56AGRA65","created_at":"2026-07-05T09:51:06.093134+00:00"},{"alias_kind":"pith_short_8","alias_value":"4WM2H6SE","created_at":"2026-07-05T09:51:06.093134+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26478","citing_title":"Efficient On-policy Visual-RL via Stochastic Decoupled Policy Gradient","ref_index":41,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB","json":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB.json","graph_json":"https://pith.science/api/pith-number/4WM2H6SE56AGRA65DSR655ICVB/graph.json","events_json":"https://pith.science/api/pith-number/4WM2H6SE56AGRA65DSR655ICVB/events.json","paper":"https://pith.science/paper/4WM2H6SE"},"agent_actions":{"view_html":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB","download_json":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB.json","view_paper":"https://pith.science/paper/4WM2H6SE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.13662&json=true","fetch_graph":"https://pith.science/api/pith-number/4WM2H6SE56AGRA65DSR655ICVB/graph.json","fetch_events":"https://pith.science/api/pith-number/4WM2H6SE56AGRA65DSR655ICVB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB/action/storage_attestation","attest_author":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB/action/author_attestation","sign_citation":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB/action/citation_signature","submit_replication":"https://pith.science/pith/4WM2H6SE56AGRA65DSR655ICVB/action/replication_record"}},"created_at":"2026-07-05T09:51:06.093134+00:00","updated_at":"2026-07-05T09:51:06.093134+00:00"}