{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:B6H6XRM6RWZSBL7F3HTOWHUWJ2","short_pith_number":"pith:B6H6XRM6","schema_version":"1.0","canonical_sha256":"0f8febc59e8db320afe5d9e6eb1e964e84131612a4db9a64ce1e1d09f1425fa1","source":{"kind":"arxiv","id":"2402.05546","version":1},"attestation_state":"computed","paper":{"title":"Offline Actor-Critic Reinforcement Learning Scales to Large Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.RO"],"primary_cat":"cs.LG","authors_text":"Abbas Abdolmaleki, Jingwei Zhang, Jost Tobias Springenberg, Martin Riedmiller, Michael Bloesch, Nicolas Heess, Oliver Groth, Philemon Brakel, Roland Hafner, Sarah Bechtle, Steven Kapturowski, Thomas Lampe","submitted_at":"2024-02-08T10:29:46Z","abstract_excerpt":"We show that offline actor-critic reinforcement learning can scale to large models - such as transformers - and follows similar scaling laws as supervised learning. We find that offline actor-critic algorithms can outperform strong, supervised, behavioral cloning baselines for multi-task training on a large dataset containing both sub-optimal and expert behavior on 132 continuous control tasks. We introduce a Perceiver-based actor-critic model and elucidate the key model features needed to make offline RL work with self- and cross-attention modules. Overall, we find that: i) simple offline act"},"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":"2402.05546","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T10:29:46Z","cross_cats_sorted":["cs.AI","cs.RO"],"title_canon_sha256":"8de4c7169bbee4c291421ca061d14b97157f8cac24a4d696ff52d333dc147abd","abstract_canon_sha256":"892498aab54099109d6e744c83c867cbc379de8aba7b1aba1d04b47a75ef025d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:42:52.262589Z","signature_b64":"I6L39MfiDeGJOgQv7AjfT9Gvp7oqtHGUMP1Ff/+r+8a87tnNf77R2Kt3dqzO412xs1RalXCZna2/ZZwRVSQXCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0f8febc59e8db320afe5d9e6eb1e964e84131612a4db9a64ce1e1d09f1425fa1","last_reissued_at":"2026-07-05T07:42:52.262145Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:42:52.262145Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Offline Actor-Critic Reinforcement Learning Scales to Large Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.RO"],"primary_cat":"cs.LG","authors_text":"Abbas Abdolmaleki, Jingwei Zhang, Jost Tobias Springenberg, Martin Riedmiller, Michael Bloesch, Nicolas Heess, Oliver Groth, Philemon Brakel, Roland Hafner, Sarah Bechtle, Steven Kapturowski, Thomas Lampe","submitted_at":"2024-02-08T10:29:46Z","abstract_excerpt":"We show that offline actor-critic reinforcement learning can scale to large models - such as transformers - and follows similar scaling laws as supervised learning. We find that offline actor-critic algorithms can outperform strong, supervised, behavioral cloning baselines for multi-task training on a large dataset containing both sub-optimal and expert behavior on 132 continuous control tasks. We introduce a Perceiver-based actor-critic model and elucidate the key model features needed to make offline RL work with self- and cross-attention modules. Overall, we find that: i) simple offline act"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05546","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/2402.05546/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":"2402.05546","created_at":"2026-07-05T07:42:52.262196+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.05546v1","created_at":"2026-07-05T07:42:52.262196+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05546","created_at":"2026-07-05T07:42:52.262196+00:00"},{"alias_kind":"pith_short_12","alias_value":"B6H6XRM6RWZS","created_at":"2026-07-05T07:42:52.262196+00:00"},{"alias_kind":"pith_short_16","alias_value":"B6H6XRM6RWZSBL7F","created_at":"2026-07-05T07:42:52.262196+00:00"},{"alias_kind":"pith_short_8","alias_value":"B6H6XRM6","created_at":"2026-07-05T07:42:52.262196+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.10289","citing_title":"Sample-Mean Anchored Thompson Sampling for Offline-to-Online Learning with Distribution Shift","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10289","citing_title":"Sample-Mean Anchored Thompson Sampling for Offline-to-Online Learning with Distribution Shift","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01663","citing_title":"Towards Efficient and Expressive Offline RL via Flow-Anchored Noise-conditioned Q-Learning","ref_index":60,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2","json":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2.json","graph_json":"https://pith.science/api/pith-number/B6H6XRM6RWZSBL7F3HTOWHUWJ2/graph.json","events_json":"https://pith.science/api/pith-number/B6H6XRM6RWZSBL7F3HTOWHUWJ2/events.json","paper":"https://pith.science/paper/B6H6XRM6"},"agent_actions":{"view_html":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2","download_json":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2.json","view_paper":"https://pith.science/paper/B6H6XRM6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.05546&json=true","fetch_graph":"https://pith.science/api/pith-number/B6H6XRM6RWZSBL7F3HTOWHUWJ2/graph.json","fetch_events":"https://pith.science/api/pith-number/B6H6XRM6RWZSBL7F3HTOWHUWJ2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2/action/storage_attestation","attest_author":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2/action/author_attestation","sign_citation":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2/action/citation_signature","submit_replication":"https://pith.science/pith/B6H6XRM6RWZSBL7F3HTOWHUWJ2/action/replication_record"}},"created_at":"2026-07-05T07:42:52.262196+00:00","updated_at":"2026-07-05T07:42:52.262196+00:00"}