{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PNTSPNDISFFJR25KM2SHUOTWR4","short_pith_number":"pith:PNTSPNDI","schema_version":"1.0","canonical_sha256":"7b6727b468914a98ebaa66a47a3a768f1f52e28759406004f90a6ccaae53fcf1","source":{"kind":"arxiv","id":"2406.04314","version":3},"attestation_state":"computed","paper":{"title":"Aesthetic Post-Training Diffusion Models from Generic Preferences with Step-by-step Preference Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bohan Chen, Ji Li, Liang Zheng, Mingxi Cheng, Shuyang Gu, Tiankai Hang, Yuhui Yuan, Zhanhao Liang","submitted_at":"2024-06-06T17:57:09Z","abstract_excerpt":"Generating visually appealing images is fundamental to modern text-to-image generation models. A potential solution to better aesthetics is direct preference optimization (DPO), which has been applied to diffusion models to improve general image quality including prompt alignment and aesthetics. Popular DPO methods propagate preference labels from clean image pairs to all the intermediate steps along the two generation trajectories. However, preference labels provided in existing datasets are blended with layout and aesthetic opinions, which would disagree with aesthetic preference. Even if ae"},"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":"2406.04314","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-06-06T17:57:09Z","cross_cats_sorted":[],"title_canon_sha256":"c83345c004f785f7a0c90f7eb3c52af4df42dbb196ad6e7302e0023fcbfe8f1f","abstract_canon_sha256":"7fa10ee218271f20271f608e5e9799819761bd212fc03633a51f92b28fe78f90"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:38:33.821667Z","signature_b64":"ZycsX8tzF3dS+g26TSrWFIxpVz5krcDKghXQiHLxQEzA4wEX24a6/+I1wq2OEK43aKMUfw0964KAq2MkQDkUBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7b6727b468914a98ebaa66a47a3a768f1f52e28759406004f90a6ccaae53fcf1","last_reissued_at":"2026-07-05T10:38:33.821170Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:38:33.821170Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Aesthetic Post-Training Diffusion Models from Generic Preferences with Step-by-step Preference Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Bohan Chen, Ji Li, Liang Zheng, Mingxi Cheng, Shuyang Gu, Tiankai Hang, Yuhui Yuan, Zhanhao Liang","submitted_at":"2024-06-06T17:57:09Z","abstract_excerpt":"Generating visually appealing images is fundamental to modern text-to-image generation models. A potential solution to better aesthetics is direct preference optimization (DPO), which has been applied to diffusion models to improve general image quality including prompt alignment and aesthetics. Popular DPO methods propagate preference labels from clean image pairs to all the intermediate steps along the two generation trajectories. However, preference labels provided in existing datasets are blended with layout and aesthetic opinions, which would disagree with aesthetic preference. Even if ae"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.04314","kind":"arxiv","version":3},"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/2406.04314/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":"2406.04314","created_at":"2026-07-05T10:38:33.821227+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.04314v3","created_at":"2026-07-05T10:38:33.821227+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.04314","created_at":"2026-07-05T10:38:33.821227+00:00"},{"alias_kind":"pith_short_12","alias_value":"PNTSPNDISFFJ","created_at":"2026-07-05T10:38:33.821227+00:00"},{"alias_kind":"pith_short_16","alias_value":"PNTSPNDISFFJR25K","created_at":"2026-07-05T10:38:33.821227+00:00"},{"alias_kind":"pith_short_8","alias_value":"PNTSPNDI","created_at":"2026-07-05T10:38:33.821227+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":8,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21498","citing_title":"Balancing Performance and Diversity in GRPO Autoregressive Text-to-Image Post-Training","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2503.12575","citing_title":"BalancedDPO: Adaptive Multi-Metric Alignment","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15855","citing_title":"Do Less, Achieve More: Do We Need Every-Step Optimization for RL Fine-tuning of Diffusion Models?","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2501.13918","citing_title":"Improving Video Generation with Human Feedback","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09433","citing_title":"Offline Preference Optimization for Rectified Flow with Noise-Tracked Pairs","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2505.07818","citing_title":"DanceGRPO: Unleashing GRPO on Visual Generation","ref_index":53,"is_internal_anchor":false},{"citing_arxiv_id":"2505.05470","citing_title":"Flow-GRPO: Training Flow Matching Models via Online RL","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04494","citing_title":"Towards General Preference Alignment: Diffusion Models at Nash Equilibrium","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4","json":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4.json","graph_json":"https://pith.science/api/pith-number/PNTSPNDISFFJR25KM2SHUOTWR4/graph.json","events_json":"https://pith.science/api/pith-number/PNTSPNDISFFJR25KM2SHUOTWR4/events.json","paper":"https://pith.science/paper/PNTSPNDI"},"agent_actions":{"view_html":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4","download_json":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4.json","view_paper":"https://pith.science/paper/PNTSPNDI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.04314&json=true","fetch_graph":"https://pith.science/api/pith-number/PNTSPNDISFFJR25KM2SHUOTWR4/graph.json","fetch_events":"https://pith.science/api/pith-number/PNTSPNDISFFJR25KM2SHUOTWR4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4/action/storage_attestation","attest_author":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4/action/author_attestation","sign_citation":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4/action/citation_signature","submit_replication":"https://pith.science/pith/PNTSPNDISFFJR25KM2SHUOTWR4/action/replication_record"}},"created_at":"2026-07-05T10:38:33.821227+00:00","updated_at":"2026-07-05T10:38:33.821227+00:00"}