{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:RZXAPTIAEFFBZYZIZEIHIQNTUJ","short_pith_number":"pith:RZXAPTIA","schema_version":"1.0","canonical_sha256":"8e6e07cd00214a1ce328c9107441b3a24c52ab4caeb358a3847bd299f83a6f5c","source":{"kind":"arxiv","id":"2505.13740","version":2},"attestation_state":"computed","paper":{"title":"Improving Compositional Generation with Diffusion Models Using Lift Scores","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chenning Yu, Sicun Gao","submitted_at":"2025-05-19T21:34:42Z","abstract_excerpt":"We introduce a novel resampling criterion using lift scores, for improving compositional generation in diffusion models. By leveraging the lift scores, we evaluate whether generated samples align with each single condition and then compose the results to determine whether the composed prompt is satisfied. Our key insight is that lift scores can be efficiently approximated using only the original diffusion model, requiring no additional training or external modules. We develop an optimized variant that achieves relatively lower computational overhead during inference while maintaining effective"},"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":"2505.13740","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-19T21:34:42Z","cross_cats_sorted":["cs.AI","cs.CV"],"title_canon_sha256":"991bb4065ab357f8f542347c83fc37321c1ab4a311f26b97aed9ebea951659cc","abstract_canon_sha256":"12f5ce0aea1076057f8d6a9a3cbdbc765ed25f5930972892cf3d62d78bb6ecdd"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:09:14.110395Z","signature_b64":"JIBehxnRHMZVE4ad4sDXJRJXp8ibSLBFXowHScc8biOPHQdrkRJbp1inwHS7qx0Djkt+gxxWlF7/zObMz1BnDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8e6e07cd00214a1ce328c9107441b3a24c52ab4caeb358a3847bd299f83a6f5c","last_reissued_at":"2026-07-05T11:09:14.109968Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:09:14.109968Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Compositional Generation with Diffusion Models Using Lift Scores","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV"],"primary_cat":"cs.LG","authors_text":"Chenning Yu, Sicun Gao","submitted_at":"2025-05-19T21:34:42Z","abstract_excerpt":"We introduce a novel resampling criterion using lift scores, for improving compositional generation in diffusion models. By leveraging the lift scores, we evaluate whether generated samples align with each single condition and then compose the results to determine whether the composed prompt is satisfied. Our key insight is that lift scores can be efficiently approximated using only the original diffusion model, requiring no additional training or external modules. We develop an optimized variant that achieves relatively lower computational overhead during inference while maintaining effective"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.13740","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/2505.13740/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":"2505.13740","created_at":"2026-07-05T11:09:14.110021+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.13740v2","created_at":"2026-07-05T11:09:14.110021+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.13740","created_at":"2026-07-05T11:09:14.110021+00:00"},{"alias_kind":"pith_short_12","alias_value":"RZXAPTIAEFFB","created_at":"2026-07-05T11:09:14.110021+00:00"},{"alias_kind":"pith_short_16","alias_value":"RZXAPTIAEFFBZYZI","created_at":"2026-07-05T11:09:14.110021+00:00"},{"alias_kind":"pith_short_8","alias_value":"RZXAPTIA","created_at":"2026-07-05T11:09:14.110021+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.23920","citing_title":"Catastrophic Compositional Generation: Why Vanilla Diffusion Models Fail to Extrapolate","ref_index":20,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ","json":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ.json","graph_json":"https://pith.science/api/pith-number/RZXAPTIAEFFBZYZIZEIHIQNTUJ/graph.json","events_json":"https://pith.science/api/pith-number/RZXAPTIAEFFBZYZIZEIHIQNTUJ/events.json","paper":"https://pith.science/paper/RZXAPTIA"},"agent_actions":{"view_html":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ","download_json":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ.json","view_paper":"https://pith.science/paper/RZXAPTIA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.13740&json=true","fetch_graph":"https://pith.science/api/pith-number/RZXAPTIAEFFBZYZIZEIHIQNTUJ/graph.json","fetch_events":"https://pith.science/api/pith-number/RZXAPTIAEFFBZYZIZEIHIQNTUJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ/action/storage_attestation","attest_author":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ/action/author_attestation","sign_citation":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ/action/citation_signature","submit_replication":"https://pith.science/pith/RZXAPTIAEFFBZYZIZEIHIQNTUJ/action/replication_record"}},"created_at":"2026-07-05T11:09:14.110021+00:00","updated_at":"2026-07-05T11:09:14.110021+00:00"}