{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TLK4TGR4ADTRWRLLDJQHOGWBP6","short_pith_number":"pith:TLK4TGR4","schema_version":"1.0","canonical_sha256":"9ad5c99a3c00e71b456b1a60771ac17f9de5222d285baf19dc6cc6d09dd06fe8","source":{"kind":"arxiv","id":"2504.18204","version":1},"attestation_state":"computed","paper":{"title":"Optimizing Multi-Round Enhanced Training in Diffusion Models for Improved Preference Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hongyang He, Jianhui Wang, Jiaqi Chen, Keqin Li, Kun Li, Miao Zhang, Ruoyu Wang, Sida Li, Tianyu Shi, Wenxin Zhang, Xinyuan Song, Xueqian Wang, Yangfan He","submitted_at":"2025-04-25T09:35:02Z","abstract_excerpt":"Generative AI has significantly changed industries by enabling text-driven image generation, yet challenges remain in achieving high-resolution outputs that align with fine-grained user preferences. Consequently, multi-round interactions are necessary to ensure the generated images meet expectations. Previous methods enhanced prompts via reward feedback but did not optimize over a multi-round dialogue dataset. In this work, we present a Visual Co-Adaptation (VCA) framework incorporating human-in-the-loop feedback, leveraging a well-trained reward model aligned with human preferences. Using a d"},"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":"2504.18204","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-25T09:35:02Z","cross_cats_sorted":[],"title_canon_sha256":"114343302d681a6ed9bdc1f2fe4bccfee555a6d4da62c44943dcab936f381434","abstract_canon_sha256":"2c34e2a70d4b89641499d5ed29ca817e4b4d209b512f7431db8c68db9bcddb29"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:54:03.678525Z","signature_b64":"byEPMgvad59BDn356vz3AhabGYPS00R3p71QwaL0P2rOnw0DDGErHoQ60R7nvUk68loEuLHg3w1Mc4N32f2PAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9ad5c99a3c00e71b456b1a60771ac17f9de5222d285baf19dc6cc6d09dd06fe8","last_reissued_at":"2026-07-05T10:54:03.678059Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:54:03.678059Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimizing Multi-Round Enhanced Training in Diffusion Models for Improved Preference Understanding","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Hongyang He, Jianhui Wang, Jiaqi Chen, Keqin Li, Kun Li, Miao Zhang, Ruoyu Wang, Sida Li, Tianyu Shi, Wenxin Zhang, Xinyuan Song, Xueqian Wang, Yangfan He","submitted_at":"2025-04-25T09:35:02Z","abstract_excerpt":"Generative AI has significantly changed industries by enabling text-driven image generation, yet challenges remain in achieving high-resolution outputs that align with fine-grained user preferences. Consequently, multi-round interactions are necessary to ensure the generated images meet expectations. Previous methods enhanced prompts via reward feedback but did not optimize over a multi-round dialogue dataset. In this work, we present a Visual Co-Adaptation (VCA) framework incorporating human-in-the-loop feedback, leveraging a well-trained reward model aligned with human preferences. Using a d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.18204","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/2504.18204/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":"2504.18204","created_at":"2026-07-05T10:54:03.678114+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.18204v1","created_at":"2026-07-05T10:54:03.678114+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.18204","created_at":"2026-07-05T10:54:03.678114+00:00"},{"alias_kind":"pith_short_12","alias_value":"TLK4TGR4ADTR","created_at":"2026-07-05T10:54:03.678114+00:00"},{"alias_kind":"pith_short_16","alias_value":"TLK4TGR4ADTRWRLL","created_at":"2026-07-05T10:54:03.678114+00:00"},{"alias_kind":"pith_short_8","alias_value":"TLK4TGR4","created_at":"2026-07-05T10:54:03.678114+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/TLK4TGR4ADTRWRLLDJQHOGWBP6","json":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6.json","graph_json":"https://pith.science/api/pith-number/TLK4TGR4ADTRWRLLDJQHOGWBP6/graph.json","events_json":"https://pith.science/api/pith-number/TLK4TGR4ADTRWRLLDJQHOGWBP6/events.json","paper":"https://pith.science/paper/TLK4TGR4"},"agent_actions":{"view_html":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6","download_json":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6.json","view_paper":"https://pith.science/paper/TLK4TGR4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.18204&json=true","fetch_graph":"https://pith.science/api/pith-number/TLK4TGR4ADTRWRLLDJQHOGWBP6/graph.json","fetch_events":"https://pith.science/api/pith-number/TLK4TGR4ADTRWRLLDJQHOGWBP6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6/action/storage_attestation","attest_author":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6/action/author_attestation","sign_citation":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6/action/citation_signature","submit_replication":"https://pith.science/pith/TLK4TGR4ADTRWRLLDJQHOGWBP6/action/replication_record"}},"created_at":"2026-07-05T10:54:03.678114+00:00","updated_at":"2026-07-05T10:54:03.678114+00:00"}