{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BP5T36X65OA2F24J76D6GIPX4T","short_pith_number":"pith:BP5T36X6","schema_version":"1.0","canonical_sha256":"0bfb3dfafeeb81a2eb89ff87e321f7e4d13898eb828c2a0795647df55a953b95","source":{"kind":"arxiv","id":"2412.07775","version":6},"attestation_state":"computed","paper":{"title":"Efficient Diversity-Preserving Diffusion Alignment via Gradient-Informed GFlowNets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Dinghuai Zhang, Tim Z. Xiao, Weiyang Liu, Yoshua Bengio, Zhen Liu","submitted_at":"2024-12-10T18:59:58Z","abstract_excerpt":"While one commonly trains large diffusion models by collecting datasets on target downstream tasks, it is often desired to align and finetune pretrained diffusion models with some reward functions that are either designed by experts or learned from small-scale datasets. Existing post-training methods for reward finetuning of diffusion models typically suffer from lack of diversity in generated samples, lack of prior preservation, and/or slow convergence in finetuning. In response to this challenge, we take inspiration from recent successes in generative flow networks (GFlowNets) and propose a "},"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":"2412.07775","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-12-10T18:59:58Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"65bbadc8a1558da35b160c264a216a9c2b6395ca6d8c40c8d7b3eb444ff422a0","abstract_canon_sha256":"0e702cc74953b4263f08cb4a6eb944c87b4387d044db1af04a7d3ed84ed0c11f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:04:42.661967Z","signature_b64":"SEieV1I08mRdeiPqBTjj0eLxGSLrPWvPAGoQf4uuiFNxI0tm/uTNP2OhZwH4J5xPmixAEi8rAv9WgQNZMGutAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0bfb3dfafeeb81a2eb89ff87e321f7e4d13898eb828c2a0795647df55a953b95","last_reissued_at":"2026-07-05T11:04:42.661504Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:04:42.661504Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Efficient Diversity-Preserving Diffusion Alignment via Gradient-Informed GFlowNets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Dinghuai Zhang, Tim Z. Xiao, Weiyang Liu, Yoshua Bengio, Zhen Liu","submitted_at":"2024-12-10T18:59:58Z","abstract_excerpt":"While one commonly trains large diffusion models by collecting datasets on target downstream tasks, it is often desired to align and finetune pretrained diffusion models with some reward functions that are either designed by experts or learned from small-scale datasets. Existing post-training methods for reward finetuning of diffusion models typically suffer from lack of diversity in generated samples, lack of prior preservation, and/or slow convergence in finetuning. In response to this challenge, we take inspiration from recent successes in generative flow networks (GFlowNets) and propose a "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.07775","kind":"arxiv","version":6},"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/2412.07775/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":"2412.07775","created_at":"2026-07-05T11:04:42.661560+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.07775v6","created_at":"2026-07-05T11:04:42.661560+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.07775","created_at":"2026-07-05T11:04:42.661560+00:00"},{"alias_kind":"pith_short_12","alias_value":"BP5T36X65OA2","created_at":"2026-07-05T11:04:42.661560+00:00"},{"alias_kind":"pith_short_16","alias_value":"BP5T36X65OA2F24J","created_at":"2026-07-05T11:04:42.661560+00:00"},{"alias_kind":"pith_short_8","alias_value":"BP5T36X6","created_at":"2026-07-05T11:04:42.661560+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T","json":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T.json","graph_json":"https://pith.science/api/pith-number/BP5T36X65OA2F24J76D6GIPX4T/graph.json","events_json":"https://pith.science/api/pith-number/BP5T36X65OA2F24J76D6GIPX4T/events.json","paper":"https://pith.science/paper/BP5T36X6"},"agent_actions":{"view_html":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T","download_json":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T.json","view_paper":"https://pith.science/paper/BP5T36X6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.07775&json=true","fetch_graph":"https://pith.science/api/pith-number/BP5T36X65OA2F24J76D6GIPX4T/graph.json","fetch_events":"https://pith.science/api/pith-number/BP5T36X65OA2F24J76D6GIPX4T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T/action/storage_attestation","attest_author":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T/action/author_attestation","sign_citation":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T/action/citation_signature","submit_replication":"https://pith.science/pith/BP5T36X65OA2F24J76D6GIPX4T/action/replication_record"}},"created_at":"2026-07-05T11:04:42.661560+00:00","updated_at":"2026-07-05T11:04:42.661560+00:00"}