{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2JNOTCPR6GUTSKZSJARF3NK5FQ","short_pith_number":"pith:2JNOTCPR","schema_version":"1.0","canonical_sha256":"d25ae989f1f1a9392b3248225db55d2c1ee25e40f83e9ad7327cc30df4a4e0dc","source":{"kind":"arxiv","id":"2406.12303","version":2},"attestation_state":"computed","paper":{"title":"Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Akio Kodaira, Chenfeng Xu, Heyang Jiang, Kurt Keutzer, Masayoshi Tomizuka, Yiheng Li","submitted_at":"2024-06-18T06:20:42Z","abstract_excerpt":"In this paper, we point out that suboptimal noise-data mapping leads to slow training of diffusion models. During diffusion training, current methods diffuse each image across the entire noise space, resulting in a mixture of all images at every point in the noise layer. We emphasize that this random mixture of noise-data mapping complicates the optimization of the denoising function in diffusion models. Drawing inspiration from the immiscibility phenomenon in physics, we propose Immiscible Diffusion, a simple and effective method to improve the random mixture of noise-data mapping. In physics"},"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":"2406.12303","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-06-18T06:20:42Z","cross_cats_sorted":[],"title_canon_sha256":"cc4b4808701a17f531fde0dc9f4c61ec77a899b77f4be2aed040d5cfe298ac9c","abstract_canon_sha256":"f3791ee209a95111e7b0383f20cdcc4b9b7392a6c994c0872aa10e9ab77c841d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:28:51.799121Z","signature_b64":"E079chdJT/y+izQeGxFABgGU/TgmetLONPi/txYqfmTW22KN6QabKV9inYHfNbNrzsuazFvXesjI6mCczIGRCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d25ae989f1f1a9392b3248225db55d2c1ee25e40f83e9ad7327cc30df4a4e0dc","last_reissued_at":"2026-07-05T09:28:51.798614Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:28:51.798614Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Immiscible Diffusion: Accelerating Diffusion Training with Noise Assignment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Akio Kodaira, Chenfeng Xu, Heyang Jiang, Kurt Keutzer, Masayoshi Tomizuka, Yiheng Li","submitted_at":"2024-06-18T06:20:42Z","abstract_excerpt":"In this paper, we point out that suboptimal noise-data mapping leads to slow training of diffusion models. During diffusion training, current methods diffuse each image across the entire noise space, resulting in a mixture of all images at every point in the noise layer. We emphasize that this random mixture of noise-data mapping complicates the optimization of the denoising function in diffusion models. Drawing inspiration from the immiscibility phenomenon in physics, we propose Immiscible Diffusion, a simple and effective method to improve the random mixture of noise-data mapping. In physics"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.12303","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/2406.12303/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":"2406.12303","created_at":"2026-07-05T09:28:51.798679+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.12303v2","created_at":"2026-07-05T09:28:51.798679+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.12303","created_at":"2026-07-05T09:28:51.798679+00:00"},{"alias_kind":"pith_short_12","alias_value":"2JNOTCPR6GUT","created_at":"2026-07-05T09:28:51.798679+00:00"},{"alias_kind":"pith_short_16","alias_value":"2JNOTCPR6GUTSKZS","created_at":"2026-07-05T09:28:51.798679+00:00"},{"alias_kind":"pith_short_8","alias_value":"2JNOTCPR","created_at":"2026-07-05T09:28:51.798679+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.06805","citing_title":"Efficient Diffusion Models: A Survey","ref_index":37,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ","json":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ.json","graph_json":"https://pith.science/api/pith-number/2JNOTCPR6GUTSKZSJARF3NK5FQ/graph.json","events_json":"https://pith.science/api/pith-number/2JNOTCPR6GUTSKZSJARF3NK5FQ/events.json","paper":"https://pith.science/paper/2JNOTCPR"},"agent_actions":{"view_html":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ","download_json":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ.json","view_paper":"https://pith.science/paper/2JNOTCPR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.12303&json=true","fetch_graph":"https://pith.science/api/pith-number/2JNOTCPR6GUTSKZSJARF3NK5FQ/graph.json","fetch_events":"https://pith.science/api/pith-number/2JNOTCPR6GUTSKZSJARF3NK5FQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ/action/storage_attestation","attest_author":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ/action/author_attestation","sign_citation":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ/action/citation_signature","submit_replication":"https://pith.science/pith/2JNOTCPR6GUTSKZSJARF3NK5FQ/action/replication_record"}},"created_at":"2026-07-05T09:28:51.798679+00:00","updated_at":"2026-07-05T09:28:51.798679+00:00"}