{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YX2RRSKOGXEZGOUNEZF6EJMHWE","short_pith_number":"pith:YX2RRSKO","schema_version":"1.0","canonical_sha256":"c5f518c94e35c9933a8d264be22587b1027241cce83994203dc5201944eacb82","source":{"kind":"arxiv","id":"2410.18082","version":2},"attestation_state":"computed","paper":{"title":"Prioritized Generative Replay","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alexei A. Efros, Kevin Frans, Pieter Abbeel, Renhao Wang, Sergey Levine","submitted_at":"2024-10-23T17:59:52Z","abstract_excerpt":"Sample-efficient online reinforcement learning often uses replay buffers to store experience for reuse when updating the value function. However, uniform replay is inefficient, since certain classes of transitions can be more relevant to learning. While prioritization of more useful samples is helpful, this strategy can also lead to overfitting, as useful samples are likely to be more rare. In this work, we instead propose a prioritized, parametric version of an agent's memory, using generative models to capture online experience. This paradigm enables (1) densification of past experience, wit"},"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":"2410.18082","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-23T17:59:52Z","cross_cats_sorted":[],"title_canon_sha256":"ec0c122c36a4a7daf17fa8aeca279b71f2bcfbb12f9167ea2fa1bb9a8fdfa1c0","abstract_canon_sha256":"30c3c93ddf63fbb134475792f4a05ae3dc93017bded846759c2a6f839c432174"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:39.821271Z","signature_b64":"2PWeJ8D1m/KduwwZaqaqZ+/x43jqyUlBEkK7tKsIQNuNHa7W5O0GOrakIRzY/so+JgiHm10TQIumHigQU1U8Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c5f518c94e35c9933a8d264be22587b1027241cce83994203dc5201944eacb82","last_reissued_at":"2026-07-05T11:00:39.820798Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:39.820798Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Prioritized Generative Replay","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alexei A. Efros, Kevin Frans, Pieter Abbeel, Renhao Wang, Sergey Levine","submitted_at":"2024-10-23T17:59:52Z","abstract_excerpt":"Sample-efficient online reinforcement learning often uses replay buffers to store experience for reuse when updating the value function. However, uniform replay is inefficient, since certain classes of transitions can be more relevant to learning. While prioritization of more useful samples is helpful, this strategy can also lead to overfitting, as useful samples are likely to be more rare. In this work, we instead propose a prioritized, parametric version of an agent's memory, using generative models to capture online experience. This paradigm enables (1) densification of past experience, wit"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.18082","kind":"arxiv","version":2},"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/2410.18082/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":"2410.18082","created_at":"2026-07-05T11:00:39.820853+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.18082v2","created_at":"2026-07-05T11:00:39.820853+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.18082","created_at":"2026-07-05T11:00:39.820853+00:00"},{"alias_kind":"pith_short_12","alias_value":"YX2RRSKOGXEZ","created_at":"2026-07-05T11:00:39.820853+00:00"},{"alias_kind":"pith_short_16","alias_value":"YX2RRSKOGXEZGOUN","created_at":"2026-07-05T11:00:39.820853+00:00"},{"alias_kind":"pith_short_8","alias_value":"YX2RRSKO","created_at":"2026-07-05T11:00:39.820853+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2506.05762","citing_title":"BiTrajDiff: Bidirectional Trajectory Generation with Diffusion Models for Offline Reinforcement Learning","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE","json":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE.json","graph_json":"https://pith.science/api/pith-number/YX2RRSKOGXEZGOUNEZF6EJMHWE/graph.json","events_json":"https://pith.science/api/pith-number/YX2RRSKOGXEZGOUNEZF6EJMHWE/events.json","paper":"https://pith.science/paper/YX2RRSKO"},"agent_actions":{"view_html":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE","download_json":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE.json","view_paper":"https://pith.science/paper/YX2RRSKO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.18082&json=true","fetch_graph":"https://pith.science/api/pith-number/YX2RRSKOGXEZGOUNEZF6EJMHWE/graph.json","fetch_events":"https://pith.science/api/pith-number/YX2RRSKOGXEZGOUNEZF6EJMHWE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE/action/storage_attestation","attest_author":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE/action/author_attestation","sign_citation":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE/action/citation_signature","submit_replication":"https://pith.science/pith/YX2RRSKOGXEZGOUNEZF6EJMHWE/action/replication_record"}},"created_at":"2026-07-05T11:00:39.820853+00:00","updated_at":"2026-07-05T11:00:39.820853+00:00"}