{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5LCXHRGZRCMBNELNZDBVQ2R77M","short_pith_number":"pith:5LCXHRGZ","schema_version":"1.0","canonical_sha256":"eac573c4d9889816916dc8c3586a3ffb03e2700b48a15dce1b59f34b7fdd4458","source":{"kind":"arxiv","id":"2407.03297","version":2},"attestation_state":"computed","paper":{"title":"Improved Noise Schedule for Diffusion Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Baining Guo, Shuyang Gu, Tiankai Hang, Xin Geng","submitted_at":"2024-07-03T17:34:55Z","abstract_excerpt":"Diffusion models have emerged as the de facto choice for generating high-quality visual signals across various domains. However, training a single model to predict noise across various levels poses significant challenges, necessitating numerous iterations and incurring significant computational costs. Various approaches, such as loss weighting strategy design and architectural refinements, have been introduced to expedite convergence and improve model performance. In this study, we propose a novel approach to design the noise schedule for enhancing the training of diffusion models. Our key ins"},"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":"2407.03297","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-07-03T17:34:55Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"822717b0adf369d9d205826409fbd8ade6c9f7503c5063a749f7a6f8569c48a5","abstract_canon_sha256":"18ba12668bfa0cc3d470353e8881fc4116b73dc20bea15693593057dff342807"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:40:58.919536Z","signature_b64":"QOONw9+2X3g0TEAd5Iqux6PC2lYn2u3fKu2SbD+IkPFOT50XK1NYWpMQj3pInmztOn7SG7REgvpuBmxyTg5iDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eac573c4d9889816916dc8c3586a3ffb03e2700b48a15dce1b59f34b7fdd4458","last_reissued_at":"2026-07-05T09:40:58.919038Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:40:58.919038Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improved Noise Schedule for Diffusion Training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Baining Guo, Shuyang Gu, Tiankai Hang, Xin Geng","submitted_at":"2024-07-03T17:34:55Z","abstract_excerpt":"Diffusion models have emerged as the de facto choice for generating high-quality visual signals across various domains. However, training a single model to predict noise across various levels poses significant challenges, necessitating numerous iterations and incurring significant computational costs. Various approaches, such as loss weighting strategy design and architectural refinements, have been introduced to expedite convergence and improve model performance. In this study, we propose a novel approach to design the noise schedule for enhancing the training of diffusion models. Our key ins"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.03297","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/2407.03297/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":"2407.03297","created_at":"2026-07-05T09:40:58.919095+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.03297v2","created_at":"2026-07-05T09:40:58.919095+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.03297","created_at":"2026-07-05T09:40:58.919095+00:00"},{"alias_kind":"pith_short_12","alias_value":"5LCXHRGZRCMB","created_at":"2026-07-05T09:40:58.919095+00:00"},{"alias_kind":"pith_short_16","alias_value":"5LCXHRGZRCMBNELN","created_at":"2026-07-05T09:40:58.919095+00:00"},{"alias_kind":"pith_short_8","alias_value":"5LCXHRGZ","created_at":"2026-07-05T09:40:58.919095+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2607.02291","citing_title":"Optimizing Visual Generative Models via Distribution-wise Rewards","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2506.01666","citing_title":"Synthesis of discrete-continuous quantum circuits with multimodal diffusion models","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11773","citing_title":"Is Monotonic Sampling Necessary in Diffusion Models?","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10790","citing_title":"Elucidating Representation Degradation Problem in Diffusion Model Training","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18966","citing_title":"Self-Improving Tabular Language Models via Iterative Reward-Guided Post-Training","ref_index":28,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M","json":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M.json","graph_json":"https://pith.science/api/pith-number/5LCXHRGZRCMBNELNZDBVQ2R77M/graph.json","events_json":"https://pith.science/api/pith-number/5LCXHRGZRCMBNELNZDBVQ2R77M/events.json","paper":"https://pith.science/paper/5LCXHRGZ"},"agent_actions":{"view_html":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M","download_json":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M.json","view_paper":"https://pith.science/paper/5LCXHRGZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.03297&json=true","fetch_graph":"https://pith.science/api/pith-number/5LCXHRGZRCMBNELNZDBVQ2R77M/graph.json","fetch_events":"https://pith.science/api/pith-number/5LCXHRGZRCMBNELNZDBVQ2R77M/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M/action/storage_attestation","attest_author":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M/action/author_attestation","sign_citation":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M/action/citation_signature","submit_replication":"https://pith.science/pith/5LCXHRGZRCMBNELNZDBVQ2R77M/action/replication_record"}},"created_at":"2026-07-05T09:40:58.919095+00:00","updated_at":"2026-07-05T09:40:58.919095+00:00"}