{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:AVH7NMFACIN5AWTO3EG5557LR7","short_pith_number":"pith:AVH7NMFA","schema_version":"1.0","canonical_sha256":"054ff6b0a0121bd05a6ed90ddef7eb8ffc3ca723af01d580b02d10388fa40c94","source":{"kind":"arxiv","id":"1912.02074","version":1},"attestation_state":"computed","paper":{"title":"AlgaeDICE: Policy Gradient from Arbitrary Experience","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bo Dai, Dale Schuurmans, Ilya Kostrikov, Lihong Li, Ofir Nachum, Yinlam Chow","submitted_at":"2019-12-04T16:06:10Z","abstract_excerpt":"In many real-world applications of reinforcement learning (RL), interactions with the environment are limited due to cost or feasibility. This presents a challenge to traditional RL algorithms since the max-return objective involves an expectation over on-policy samples. We introduce a new formulation of max-return optimization that allows the problem to be re-expressed by an expectation over an arbitrary behavior-agnostic and off-policy data distribution. We first derive this result by considering a regularized version of the dual max-return objective before extending our findings to unregula"},"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":"1912.02074","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2019-12-04T16:06:10Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9ec97c19b6346ff6625df39d258d4accc4c656f288ae4467c9ffa66b3e8cf2b5","abstract_canon_sha256":"e37bcc78139103d45a27227719d963960adef634ac7045d03e21429266ca7029"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:23:56.651417Z","signature_b64":"HKqQAVSlaDo5IklpZO8qQFPqvIQCQgGUOP4kJi8pxTObC8ZGoKAknoRBzeN90xeODTbSvpPkkgcG5bPdnzrCBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"054ff6b0a0121bd05a6ed90ddef7eb8ffc3ca723af01d580b02d10388fa40c94","last_reissued_at":"2026-07-05T00:23:56.650905Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:23:56.650905Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"AlgaeDICE: Policy Gradient from Arbitrary Experience","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Bo Dai, Dale Schuurmans, Ilya Kostrikov, Lihong Li, Ofir Nachum, Yinlam Chow","submitted_at":"2019-12-04T16:06:10Z","abstract_excerpt":"In many real-world applications of reinforcement learning (RL), interactions with the environment are limited due to cost or feasibility. This presents a challenge to traditional RL algorithms since the max-return objective involves an expectation over on-policy samples. We introduce a new formulation of max-return optimization that allows the problem to be re-expressed by an expectation over an arbitrary behavior-agnostic and off-policy data distribution. We first derive this result by considering a regularized version of the dual max-return objective before extending our findings to unregula"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1912.02074","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/1912.02074/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":"1912.02074","created_at":"2026-07-05T00:23:56.650966+00:00"},{"alias_kind":"arxiv_version","alias_value":"1912.02074v1","created_at":"2026-07-05T00:23:56.650966+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1912.02074","created_at":"2026-07-05T00:23:56.650966+00:00"},{"alias_kind":"pith_short_12","alias_value":"AVH7NMFACIN5","created_at":"2026-07-05T00:23:56.650966+00:00"},{"alias_kind":"pith_short_16","alias_value":"AVH7NMFACIN5AWTO","created_at":"2026-07-05T00:23:56.650966+00:00"},{"alias_kind":"pith_short_8","alias_value":"AVH7NMFA","created_at":"2026-07-05T00:23:56.650966+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05375","citing_title":"Fitted Occupancy-Ratio Evaluation without Bellman Completeness","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2408.08812","citing_title":"TRAM: Test-Time Risk Adaptation with Mixture of Agents","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2510.01479","citing_title":"Density-Ratio Weighted Behavioral Cloning: Learning Control Policies from Corrupted Datasets","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2512.23927","citing_title":"Stationary Reweighting Yields Local Convergence of Soft Fitted Q-Iteration","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2512.23805","citing_title":"Fitted $Q$ Evaluation Without Bellman Completeness via Stationary Weighting","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2210.00030","citing_title":"VIP: Towards Universal Visual Reward and Representation via Value-Implicit Pre-Training","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2004.07219","citing_title":"D4RL: Datasets for Deep Data-Driven Reinforcement Learning","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2005.01643","citing_title":"Offline Reinforcement Learning: Tutorial, Review, and Perspectives on Open Problems","ref_index":188,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7","json":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7.json","graph_json":"https://pith.science/api/pith-number/AVH7NMFACIN5AWTO3EG5557LR7/graph.json","events_json":"https://pith.science/api/pith-number/AVH7NMFACIN5AWTO3EG5557LR7/events.json","paper":"https://pith.science/paper/AVH7NMFA"},"agent_actions":{"view_html":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7","download_json":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7.json","view_paper":"https://pith.science/paper/AVH7NMFA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1912.02074&json=true","fetch_graph":"https://pith.science/api/pith-number/AVH7NMFACIN5AWTO3EG5557LR7/graph.json","fetch_events":"https://pith.science/api/pith-number/AVH7NMFACIN5AWTO3EG5557LR7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7/action/storage_attestation","attest_author":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7/action/author_attestation","sign_citation":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7/action/citation_signature","submit_replication":"https://pith.science/pith/AVH7NMFACIN5AWTO3EG5557LR7/action/replication_record"}},"created_at":"2026-07-05T00:23:56.650966+00:00","updated_at":"2026-07-05T00:23:56.650966+00:00"}