{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:I7OCA2KT7IGBFW5AJYXB364725","short_pith_number":"pith:I7OCA2KT","schema_version":"1.0","canonical_sha256":"47dc206953fa0c12dba04e2e1dfb9fd7651d44dbea8b6a7eb5be831079c425fb","source":{"kind":"arxiv","id":"2512.19311","version":2},"attestation_state":"computed","paper":{"title":"MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Fu-Yun Wang, Haoyuan Xia, Hui Li, Jiayue Lyu, Jingdong Wang, Kaihui Cheng, Siyu Zhu","submitted_at":"2025-12-22T12:00:12Z","abstract_excerpt":"This paper studies the training-testing discrepancy (a.k.a. exposure bias) problem for improving the diffusion models. During training, the input of a prediction network at one training timestep is the corresponding ground-truth noisy data that is an interpolation of the noise and the data, and during testing, the input is the generated noisy data. We present a novel training approach, named MixFlow, for improving the performance. Our approach is motivated by the Slow Flow phenomenon: the ground-truth interpolation that is the nearest to the generated noisy data at a given sampling timestep is"},"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":"2512.19311","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-12-22T12:00:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"6ff52e02489050d43fc6edba88c901dc03c8cd8ffe1d7bb62c77bb09f84e27da","abstract_canon_sha256":"17a490131de405ada8ca7b663d313923686311b2cea744f306d390457cbe064c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-14T01:20:50.109149Z","signature_b64":"8QemhT6Nbc62pJA4DRd7RlkdqcDX20QAE6neJJByJzcsAdJxipb5+x2EqTxRkrN8vZPh5EHHXwo4lbbqrMz4Cg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"47dc206953fa0c12dba04e2e1dfb9fd7651d44dbea8b6a7eb5be831079c425fb","last_reissued_at":"2026-07-14T01:20:50.108262Z","signature_status":"signed_v1","first_computed_at":"2026-07-14T01:20:50.108262Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MixFlow Training: Alleviating Exposure Bias with Slowed Interpolation Mixture","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Fu-Yun Wang, Haoyuan Xia, Hui Li, Jiayue Lyu, Jingdong Wang, Kaihui Cheng, Siyu Zhu","submitted_at":"2025-12-22T12:00:12Z","abstract_excerpt":"This paper studies the training-testing discrepancy (a.k.a. exposure bias) problem for improving the diffusion models. During training, the input of a prediction network at one training timestep is the corresponding ground-truth noisy data that is an interpolation of the noise and the data, and during testing, the input is the generated noisy data. We present a novel training approach, named MixFlow, for improving the performance. Our approach is motivated by the Slow Flow phenomenon: the ground-truth interpolation that is the nearest to the generated noisy data at a given sampling timestep is"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.19311","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/2512.19311/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":"2512.19311","created_at":"2026-07-14T01:20:50.108690+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.19311v2","created_at":"2026-07-14T01:20:50.108690+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.19311","created_at":"2026-07-14T01:20:50.108690+00:00"},{"alias_kind":"pith_short_12","alias_value":"I7OCA2KT7IGB","created_at":"2026-07-14T01:20:50.108690+00:00"},{"alias_kind":"pith_short_16","alias_value":"I7OCA2KT7IGBFW5A","created_at":"2026-07-14T01:20:50.108690+00:00"},{"alias_kind":"pith_short_8","alias_value":"I7OCA2KT","created_at":"2026-07-14T01:20:50.108690+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/I7OCA2KT7IGBFW5AJYXB364725","json":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725.json","graph_json":"https://pith.science/api/pith-number/I7OCA2KT7IGBFW5AJYXB364725/graph.json","events_json":"https://pith.science/api/pith-number/I7OCA2KT7IGBFW5AJYXB364725/events.json","paper":"https://pith.science/paper/I7OCA2KT"},"agent_actions":{"view_html":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725","download_json":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725.json","view_paper":"https://pith.science/paper/I7OCA2KT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.19311&json=true","fetch_graph":"https://pith.science/api/pith-number/I7OCA2KT7IGBFW5AJYXB364725/graph.json","fetch_events":"https://pith.science/api/pith-number/I7OCA2KT7IGBFW5AJYXB364725/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725/action/storage_attestation","attest_author":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725/action/author_attestation","sign_citation":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725/action/citation_signature","submit_replication":"https://pith.science/pith/I7OCA2KT7IGBFW5AJYXB364725/action/replication_record"}},"created_at":"2026-07-14T01:20:50.108690+00:00","updated_at":"2026-07-14T01:20:50.108690+00:00"}