{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:HHEYODLFGTI7RCCBMEV5XFPAXE","short_pith_number":"pith:HHEYODLF","schema_version":"1.0","canonical_sha256":"39c9870d6534d1f88841612bdb95e0b932b67944eefc37cdf1b0b8af9cb122f3","source":{"kind":"arxiv","id":"2304.00046","version":1},"attestation_state":"computed","paper":{"title":"Accelerating exploration and representation learning with offline pre-training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ankit Anand, Bogdan Mazoure, Doina Precup, Jake Bruce, Rob Fergus","submitted_at":"2023-03-31T18:03:30Z","abstract_excerpt":"Sequential decision-making agents struggle with long horizon tasks, since solving them requires multi-step reasoning. Most reinforcement learning (RL) algorithms address this challenge by improved credit assignment, introducing memory capability, altering the agent's intrinsic motivation (i.e. exploration) or its worldview (i.e. knowledge representation). Many of these components could be learned from offline data. In this work, we follow the hypothesis that exploration and representation learning can be improved by separately learning two different models from a single offline dataset. We sho"},"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":"2304.00046","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-03-31T18:03:30Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8e28d9dfa7471fac05f9f2c5583e95633fc971a8892c96d993eb16eb744187b5","abstract_canon_sha256":"68f560f824572229f9f67bf8b41c130a1e0f60e43764361a32c98ed8ab345968"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:56:58.002713Z","signature_b64":"vMUQ13RlZE3PRhNiTjTaiotdsDB15G/6UHqv90BjSvK9q5fmAKTY4b3rh4LisgannWnJ1R7MvNjwyUdYTUztCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"39c9870d6534d1f88841612bdb95e0b932b67944eefc37cdf1b0b8af9cb122f3","last_reissued_at":"2026-07-05T05:56:58.002214Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:56:58.002214Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accelerating exploration and representation learning with offline pre-training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Ankit Anand, Bogdan Mazoure, Doina Precup, Jake Bruce, Rob Fergus","submitted_at":"2023-03-31T18:03:30Z","abstract_excerpt":"Sequential decision-making agents struggle with long horizon tasks, since solving them requires multi-step reasoning. Most reinforcement learning (RL) algorithms address this challenge by improved credit assignment, introducing memory capability, altering the agent's intrinsic motivation (i.e. exploration) or its worldview (i.e. knowledge representation). Many of these components could be learned from offline data. In this work, we follow the hypothesis that exploration and representation learning can be improved by separately learning two different models from a single offline dataset. We sho"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2304.00046","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/2304.00046/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":"2304.00046","created_at":"2026-07-05T05:56:58.002269+00:00"},{"alias_kind":"arxiv_version","alias_value":"2304.00046v1","created_at":"2026-07-05T05:56:58.002269+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2304.00046","created_at":"2026-07-05T05:56:58.002269+00:00"},{"alias_kind":"pith_short_12","alias_value":"HHEYODLFGTI7","created_at":"2026-07-05T05:56:58.002269+00:00"},{"alias_kind":"pith_short_16","alias_value":"HHEYODLFGTI7RCCB","created_at":"2026-07-05T05:56:58.002269+00:00"},{"alias_kind":"pith_short_8","alias_value":"HHEYODLF","created_at":"2026-07-05T05:56:58.002269+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2501.08925","citing_title":"Disentangling Exploration of Large Language Models by Optimal Exploitation","ref_index":35,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE","json":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE.json","graph_json":"https://pith.science/api/pith-number/HHEYODLFGTI7RCCBMEV5XFPAXE/graph.json","events_json":"https://pith.science/api/pith-number/HHEYODLFGTI7RCCBMEV5XFPAXE/events.json","paper":"https://pith.science/paper/HHEYODLF"},"agent_actions":{"view_html":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE","download_json":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE.json","view_paper":"https://pith.science/paper/HHEYODLF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2304.00046&json=true","fetch_graph":"https://pith.science/api/pith-number/HHEYODLFGTI7RCCBMEV5XFPAXE/graph.json","fetch_events":"https://pith.science/api/pith-number/HHEYODLFGTI7RCCBMEV5XFPAXE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE/action/storage_attestation","attest_author":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE/action/author_attestation","sign_citation":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE/action/citation_signature","submit_replication":"https://pith.science/pith/HHEYODLFGTI7RCCBMEV5XFPAXE/action/replication_record"}},"created_at":"2026-07-05T05:56:58.002269+00:00","updated_at":"2026-07-05T05:56:58.002269+00:00"}