{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4GG3CGWNSYTRRTGM7RTUO5UMXE","short_pith_number":"pith:4GG3CGWN","schema_version":"1.0","canonical_sha256":"e18db11acd962718ccccfc6747768cb93c0aa9acbb138af607e196b8f257b707","source":{"kind":"arxiv","id":"2605.28900","version":1},"attestation_state":"computed","paper":{"title":"Spectral Guidance for Flexible and Efficient Control of Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenyan Xiong, Gabriel Moreira, Jo\\~ao Paulo Costeira, Manuel Marques","submitted_at":"2026-05-27T15:11:09Z","abstract_excerpt":"We introduce Spectral Guidance, a framework for controlling diffusion models by leveraging the intrinsic geometry of the generative process. As data is progressively corrupted by noise, only a small number of features remain informative for control. We characterize them as the singular functions of a conditional expectation operator and show that they can be learned via a self-supervised objective. Once recovered, this basis enables the projection of arbitrary guidance signals, such as labels, CLIP embeddings, or masks, directly onto the sampling trajectory. This approach allows for stable, hi"},"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":"2605.28900","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T15:11:09Z","cross_cats_sorted":[],"title_canon_sha256":"3b26273ec8136c6d372da10f5c1e19ec0802cf9da37c9a887bb0b6b1e319c73e","abstract_canon_sha256":"f05395200be0e452aa174c0cc3b7a9c73345fad926c50c01994769c92fb0a422"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T00:04:15.818134Z","signature_b64":"ruD0BXUWEGVD6Oy/3H27n3C6Y1xkwNMZhN4/G3r9mWYc9qFZW14HtXFNaiH55L9ztA6R6Z0EZWFYaFNnWdsMDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e18db11acd962718ccccfc6747768cb93c0aa9acbb138af607e196b8f257b707","last_reissued_at":"2026-05-29T00:04:15.817677Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T00:04:15.817677Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Spectral Guidance for Flexible and Efficient Control of Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenyan Xiong, Gabriel Moreira, Jo\\~ao Paulo Costeira, Manuel Marques","submitted_at":"2026-05-27T15:11:09Z","abstract_excerpt":"We introduce Spectral Guidance, a framework for controlling diffusion models by leveraging the intrinsic geometry of the generative process. As data is progressively corrupted by noise, only a small number of features remain informative for control. We characterize them as the singular functions of a conditional expectation operator and show that they can be learned via a self-supervised objective. Once recovered, this basis enables the projection of arbitrary guidance signals, such as labels, CLIP embeddings, or masks, directly onto the sampling trajectory. This approach allows for stable, hi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28900","kind":"arxiv","version":1},"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/2605.28900/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":"2605.28900","created_at":"2026-05-29T00:04:15.817741+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28900v1","created_at":"2026-05-29T00:04:15.817741+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28900","created_at":"2026-05-29T00:04:15.817741+00:00"},{"alias_kind":"pith_short_12","alias_value":"4GG3CGWNSYTR","created_at":"2026-05-29T00:04:15.817741+00:00"},{"alias_kind":"pith_short_16","alias_value":"4GG3CGWNSYTRRTGM","created_at":"2026-05-29T00:04:15.817741+00:00"},{"alias_kind":"pith_short_8","alias_value":"4GG3CGWN","created_at":"2026-05-29T00:04:15.817741+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/4GG3CGWNSYTRRTGM7RTUO5UMXE","json":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE.json","graph_json":"https://pith.science/api/pith-number/4GG3CGWNSYTRRTGM7RTUO5UMXE/graph.json","events_json":"https://pith.science/api/pith-number/4GG3CGWNSYTRRTGM7RTUO5UMXE/events.json","paper":"https://pith.science/paper/4GG3CGWN"},"agent_actions":{"view_html":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE","download_json":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE.json","view_paper":"https://pith.science/paper/4GG3CGWN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28900&json=true","fetch_graph":"https://pith.science/api/pith-number/4GG3CGWNSYTRRTGM7RTUO5UMXE/graph.json","fetch_events":"https://pith.science/api/pith-number/4GG3CGWNSYTRRTGM7RTUO5UMXE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE/action/storage_attestation","attest_author":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE/action/author_attestation","sign_citation":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE/action/citation_signature","submit_replication":"https://pith.science/pith/4GG3CGWNSYTRRTGM7RTUO5UMXE/action/replication_record"}},"created_at":"2026-05-29T00:04:15.817741+00:00","updated_at":"2026-05-29T00:04:15.817741+00:00"}