{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:Q6UQKEBR47PZFLC6AR5BCKRIOC","short_pith_number":"pith:Q6UQKEBR","schema_version":"1.0","canonical_sha256":"87a9051031e7df92ac5e047a112a2870a7919b29190028df0b3e46f962bf13c0","source":{"kind":"arxiv","id":"2311.14097","version":3},"attestation_state":"computed","paper":{"title":"ACT-Diffusion: Efficient Adversarial Consistency Training for One-step Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fei Kong, Hao Cheng, Hengtao Shen, Jinhao Duan, Kaidi Xu, Lichao Sun, Renjing Xu, Xiaofeng Zhu, Xiaoshuang Shi","submitted_at":"2023-11-23T16:49:06Z","abstract_excerpt":"Though diffusion models excel in image generation, their step-by-step denoising leads to slow generation speeds. Consistency training addresses this issue with single-step sampling but often produces lower-quality generations and requires high training costs. In this paper, we show that optimizing consistency training loss minimizes the Wasserstein distance between target and generated distributions. As timestep increases, the upper bound accumulates previous consistency training losses. Therefore, larger batch sizes are needed to reduce both current and accumulated losses. We propose Adversar"},"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":"2311.14097","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-11-23T16:49:06Z","cross_cats_sorted":[],"title_canon_sha256":"6e675d5bbec56e4544b994e3065483c1735bce9c6568d8ec8136768b3298c8b0","abstract_canon_sha256":"02d48a4483ba57d20540218a9e938c6822f6d60e6de6e6c3707fce96176b6a06"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:01:55.876378Z","signature_b64":"/VyZR0/Kir0XeVbdXFUo/hXObMekAs5z9G6DsFls2CYD5FLmjqFPC8/LXl/SaX//p9WQ9QnGDdvQzbt+hdA4Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"87a9051031e7df92ac5e047a112a2870a7919b29190028df0b3e46f962bf13c0","last_reissued_at":"2026-07-05T08:01:55.875838Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:01:55.875838Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ACT-Diffusion: Efficient Adversarial Consistency Training for One-step Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Fei Kong, Hao Cheng, Hengtao Shen, Jinhao Duan, Kaidi Xu, Lichao Sun, Renjing Xu, Xiaofeng Zhu, Xiaoshuang Shi","submitted_at":"2023-11-23T16:49:06Z","abstract_excerpt":"Though diffusion models excel in image generation, their step-by-step denoising leads to slow generation speeds. Consistency training addresses this issue with single-step sampling but often produces lower-quality generations and requires high training costs. In this paper, we show that optimizing consistency training loss minimizes the Wasserstein distance between target and generated distributions. As timestep increases, the upper bound accumulates previous consistency training losses. Therefore, larger batch sizes are needed to reduce both current and accumulated losses. We propose Adversar"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.14097","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/2311.14097/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":"2311.14097","created_at":"2026-07-05T08:01:55.875898+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.14097v3","created_at":"2026-07-05T08:01:55.875898+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.14097","created_at":"2026-07-05T08:01:55.875898+00:00"},{"alias_kind":"pith_short_12","alias_value":"Q6UQKEBR47PZ","created_at":"2026-07-05T08:01:55.875898+00:00"},{"alias_kind":"pith_short_16","alias_value":"Q6UQKEBR47PZFLC6","created_at":"2026-07-05T08:01:55.875898+00:00"},{"alias_kind":"pith_short_8","alias_value":"Q6UQKEBR","created_at":"2026-07-05T08:01:55.875898+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.01441","citing_title":"Improved Training Technique for Latent Consistency Models","ref_index":11,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC","json":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC.json","graph_json":"https://pith.science/api/pith-number/Q6UQKEBR47PZFLC6AR5BCKRIOC/graph.json","events_json":"https://pith.science/api/pith-number/Q6UQKEBR47PZFLC6AR5BCKRIOC/events.json","paper":"https://pith.science/paper/Q6UQKEBR"},"agent_actions":{"view_html":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC","download_json":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC.json","view_paper":"https://pith.science/paper/Q6UQKEBR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.14097&json=true","fetch_graph":"https://pith.science/api/pith-number/Q6UQKEBR47PZFLC6AR5BCKRIOC/graph.json","fetch_events":"https://pith.science/api/pith-number/Q6UQKEBR47PZFLC6AR5BCKRIOC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC/action/storage_attestation","attest_author":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC/action/author_attestation","sign_citation":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC/action/citation_signature","submit_replication":"https://pith.science/pith/Q6UQKEBR47PZFLC6AR5BCKRIOC/action/replication_record"}},"created_at":"2026-07-05T08:01:55.875898+00:00","updated_at":"2026-07-05T08:01:55.875898+00:00"}