{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:I2QLG7LCKME6CBNFWKIAY25I66","short_pith_number":"pith:I2QLG7LC","schema_version":"1.0","canonical_sha256":"46a0b37d625309e105a5b2900c6ba8f7a112d02b60407b3b59d7f8e40f8c656d","source":{"kind":"arxiv","id":"2602.18647","version":2},"attestation_state":"computed","paper":{"title":"Noise Scheduling as Information-Guided Allocation in Diffusion Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Bac Nguyen, Chieh-Hsin Lai, Dejan Stancevic, Gabriel Raya, Georgios Batzolis, Luca Ambrogioni, Naoki Murata, Yuhta Takida, Yuki Mitsufuji","submitted_at":"2026-02-20T22:45:01Z","abstract_excerpt":"We introduce InfoNoise, an online adaptive noise schedule for diffusion training that reallocates optimization effort toward noise levels where denoising is most informative. Together with loss weighting, a noise schedule induces an effective allocation across denoising problems, often fixed before informative noise levels are known. InfoNoise makes this allocation data-adaptive by estimating a conditional-entropy-rate profile from denoising losses during training, without auxiliary models or offline search. Through I--MMSE, this profile identifies where noisy observations rapidly reduce uncer"},"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":"2602.18647","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-02-20T22:45:01Z","cross_cats_sorted":["cs.AI","cs.CV","cs.IT","math.IT"],"title_canon_sha256":"24493bf0a4da02ad67b629a45d31a32b64c2c2aa70fe90fb2cca16ad8b32be4a","abstract_canon_sha256":"2900da899b6e59294bcc22e92fe20d671b0bca4d1e892dce3fcb87ccf72809cd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:11.952412Z","signature_b64":"E9LC9PTHrvnKn7JsYCbkavYWdecc7hDLR+kfn94OO+qSLb0WRPKUNLqiwCJBFXZe8lFE6rPh+GqRKsFKbp60Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46a0b37d625309e105a5b2900c6ba8f7a112d02b60407b3b59d7f8e40f8c656d","last_reissued_at":"2026-05-28T01:05:11.951873Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:11.951873Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Noise Scheduling as Information-Guided Allocation in Diffusion Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","cs.IT","math.IT"],"primary_cat":"cs.LG","authors_text":"Bac Nguyen, Chieh-Hsin Lai, Dejan Stancevic, Gabriel Raya, Georgios Batzolis, Luca Ambrogioni, Naoki Murata, Yuhta Takida, Yuki Mitsufuji","submitted_at":"2026-02-20T22:45:01Z","abstract_excerpt":"We introduce InfoNoise, an online adaptive noise schedule for diffusion training that reallocates optimization effort toward noise levels where denoising is most informative. Together with loss weighting, a noise schedule induces an effective allocation across denoising problems, often fixed before informative noise levels are known. InfoNoise makes this allocation data-adaptive by estimating a conditional-entropy-rate profile from denoising losses during training, without auxiliary models or offline search. Through I--MMSE, this profile identifies where noisy observations rapidly reduce uncer"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.18647","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/2602.18647/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":"2602.18647","created_at":"2026-05-28T01:05:11.951939+00:00"},{"alias_kind":"arxiv_version","alias_value":"2602.18647v2","created_at":"2026-05-28T01:05:11.951939+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.18647","created_at":"2026-05-28T01:05:11.951939+00:00"},{"alias_kind":"pith_short_12","alias_value":"I2QLG7LCKME6","created_at":"2026-05-28T01:05:11.951939+00:00"},{"alias_kind":"pith_short_16","alias_value":"I2QLG7LCKME6CBNF","created_at":"2026-05-28T01:05:11.951939+00:00"},{"alias_kind":"pith_short_8","alias_value":"I2QLG7LC","created_at":"2026-05-28T01:05:11.951939+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.08144","citing_title":"NoiseRater: Meta-Learned Noise Valuation for Diffusion Model Training","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07013","citing_title":"Towards Closing the Autoregressive Gap in Language Modeling via Entropy-Gated Continuous Bitstream Diffusion","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66","json":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66.json","graph_json":"https://pith.science/api/pith-number/I2QLG7LCKME6CBNFWKIAY25I66/graph.json","events_json":"https://pith.science/api/pith-number/I2QLG7LCKME6CBNFWKIAY25I66/events.json","paper":"https://pith.science/paper/I2QLG7LC"},"agent_actions":{"view_html":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66","download_json":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66.json","view_paper":"https://pith.science/paper/I2QLG7LC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2602.18647&json=true","fetch_graph":"https://pith.science/api/pith-number/I2QLG7LCKME6CBNFWKIAY25I66/graph.json","fetch_events":"https://pith.science/api/pith-number/I2QLG7LCKME6CBNFWKIAY25I66/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66/action/timestamp_anchor","attest_storage":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66/action/storage_attestation","attest_author":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66/action/author_attestation","sign_citation":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66/action/citation_signature","submit_replication":"https://pith.science/pith/I2QLG7LCKME6CBNFWKIAY25I66/action/replication_record"}},"created_at":"2026-05-28T01:05:11.951939+00:00","updated_at":"2026-05-28T01:05:11.951939+00:00"}