{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YFRL6Z2I4BBLNU7VQL54YPRN32","short_pith_number":"pith:YFRL6Z2I","schema_version":"1.0","canonical_sha256":"c162bf6748e042b6d3f582fbcc3e2dde8f43c113f2564caafa35baafed399f1f","source":{"kind":"arxiv","id":"2509.01632","version":1},"attestation_state":"computed","paper":{"title":"Relative Trajectory Balance is equivalent to Trust-PCL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Doina Precup, Padideh Nouri, Tristan Deleu, Yoshua Bengio","submitted_at":"2025-09-01T17:17:25Z","abstract_excerpt":"Recent progress in generative modeling has highlighted the importance of Reinforcement Learning (RL) for fine-tuning, with KL-regularized methods in particular proving to be highly effective for both autoregressive and diffusion models. Complementing this line of work, the Relative Trajectory Balance (RTB) objective was recently introduced in the context of Generative Flow Networks (GFlowNets) to serve the same role of improving fine-tuning in sequential generative models. Building on prior work linking GFlowNets and maximum-entropy RL, we establish in this paper an equivalence between RTB and"},"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":"2509.01632","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-09-01T17:17:25Z","cross_cats_sorted":[],"title_canon_sha256":"c3d30c43090ce3c7ad86f95ebf8ff7e751a13ad769de857b41f4bde9487d8782","abstract_canon_sha256":"6c133cf4772e203eabcedf49c580dfaeca933fde2b18133328246a2addaabeac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:03:04.937581Z","signature_b64":"CLUWASExA3KAXls9yhV4TWrs4nD6NgrqXV6gVKXHVqU01oulZxJS0ykaUilYLOR3aIoHYz8tdww5RsIO09V7Dg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c162bf6748e042b6d3f582fbcc3e2dde8f43c113f2564caafa35baafed399f1f","last_reissued_at":"2026-07-05T12:03:04.936978Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:03:04.936978Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Relative Trajectory Balance is equivalent to Trust-PCL","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Doina Precup, Padideh Nouri, Tristan Deleu, Yoshua Bengio","submitted_at":"2025-09-01T17:17:25Z","abstract_excerpt":"Recent progress in generative modeling has highlighted the importance of Reinforcement Learning (RL) for fine-tuning, with KL-regularized methods in particular proving to be highly effective for both autoregressive and diffusion models. Complementing this line of work, the Relative Trajectory Balance (RTB) objective was recently introduced in the context of Generative Flow Networks (GFlowNets) to serve the same role of improving fine-tuning in sequential generative models. Building on prior work linking GFlowNets and maximum-entropy RL, we establish in this paper an equivalence between RTB and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.01632","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/2509.01632/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":"2509.01632","created_at":"2026-07-05T12:03:04.937053+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.01632v1","created_at":"2026-07-05T12:03:04.937053+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.01632","created_at":"2026-07-05T12:03:04.937053+00:00"},{"alias_kind":"pith_short_12","alias_value":"YFRL6Z2I4BBL","created_at":"2026-07-05T12:03:04.937053+00:00"},{"alias_kind":"pith_short_16","alias_value":"YFRL6Z2I4BBLNU7V","created_at":"2026-07-05T12:03:04.937053+00:00"},{"alias_kind":"pith_short_8","alias_value":"YFRL6Z2I","created_at":"2026-07-05T12:03:04.937053+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26552","citing_title":"Aligning Few-Step Generative Models by Amortizing Sample-based Variational Inference","ref_index":15,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32","json":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32.json","graph_json":"https://pith.science/api/pith-number/YFRL6Z2I4BBLNU7VQL54YPRN32/graph.json","events_json":"https://pith.science/api/pith-number/YFRL6Z2I4BBLNU7VQL54YPRN32/events.json","paper":"https://pith.science/paper/YFRL6Z2I"},"agent_actions":{"view_html":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32","download_json":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32.json","view_paper":"https://pith.science/paper/YFRL6Z2I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.01632&json=true","fetch_graph":"https://pith.science/api/pith-number/YFRL6Z2I4BBLNU7VQL54YPRN32/graph.json","fetch_events":"https://pith.science/api/pith-number/YFRL6Z2I4BBLNU7VQL54YPRN32/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32/action/storage_attestation","attest_author":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32/action/author_attestation","sign_citation":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32/action/citation_signature","submit_replication":"https://pith.science/pith/YFRL6Z2I4BBLNU7VQL54YPRN32/action/replication_record"}},"created_at":"2026-07-05T12:03:04.937053+00:00","updated_at":"2026-07-05T12:03:04.937053+00:00"}