{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:57VKYP4VZTBY343UO7DG2TWTYX","short_pith_number":"pith:57VKYP4V","schema_version":"1.0","canonical_sha256":"efeaac3f95ccc38df37477c66d4ed3c5efdd93066fb43614a7e37b95588e4156","source":{"kind":"arxiv","id":"2508.09156","version":3},"attestation_state":"computed","paper":{"title":"Physics-Constrained Fine-Tuning of Flow-Matching Models for Generation and Inverse Problems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.AP"],"primary_cat":"cs.LG","authors_text":"Andrew B. Duncan, Jan Tauberschmidt, Sebastian J. Vollmer, Sophie Fellenz","submitted_at":"2025-08-05T09:32:04Z","abstract_excerpt":"We present a framework for fine-tuning flow-matching generative models to enforce physical constraints and solve inverse problems in scientific systems. Starting from a model trained on low-fidelity or observational data, we apply a differentiable post-training procedure that minimizes weak-form residuals of governing partial differential equations (PDEs), promoting physical consistency and adherence to boundary conditions without distorting the underlying learned distribution. To infer unknown physical inputs, such as source terms, material parameters, or boundary data, we augment the generat"},"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":"2508.09156","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-08-05T09:32:04Z","cross_cats_sorted":["cs.AI","stat.AP"],"title_canon_sha256":"dcae5960f3249aa27affd36522bc1dedabe200cc546796f9f14833f3fe02a143","abstract_canon_sha256":"060cee2c7ce36067068defa6ba58ac654709101b7ec138fe9dd8e13a86ac3040"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T01:17:40.927391Z","signature_b64":"RC38Q6crOVnRXDTCGoK3Iv9oI86x+ByZYI8H0rLV8cxYIaL8rG4waP9pCiYMsVvUYQpo/ri8s6P5DHmaSv42DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"efeaac3f95ccc38df37477c66d4ed3c5efdd93066fb43614a7e37b95588e4156","last_reissued_at":"2026-07-01T01:17:40.926875Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T01:17:40.926875Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Physics-Constrained Fine-Tuning of Flow-Matching Models for Generation and Inverse Problems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","stat.AP"],"primary_cat":"cs.LG","authors_text":"Andrew B. Duncan, Jan Tauberschmidt, Sebastian J. Vollmer, Sophie Fellenz","submitted_at":"2025-08-05T09:32:04Z","abstract_excerpt":"We present a framework for fine-tuning flow-matching generative models to enforce physical constraints and solve inverse problems in scientific systems. Starting from a model trained on low-fidelity or observational data, we apply a differentiable post-training procedure that minimizes weak-form residuals of governing partial differential equations (PDEs), promoting physical consistency and adherence to boundary conditions without distorting the underlying learned distribution. To infer unknown physical inputs, such as source terms, material parameters, or boundary data, we augment the generat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.09156","kind":"arxiv","version":3},"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/2508.09156/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":"2508.09156","created_at":"2026-07-01T01:17:40.926935+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.09156v3","created_at":"2026-07-01T01:17:40.926935+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.09156","created_at":"2026-07-01T01:17:40.926935+00:00"},{"alias_kind":"pith_short_12","alias_value":"57VKYP4VZTBY","created_at":"2026-07-01T01:17:40.926935+00:00"},{"alias_kind":"pith_short_16","alias_value":"57VKYP4VZTBY343U","created_at":"2026-07-01T01:17:40.926935+00:00"},{"alias_kind":"pith_short_8","alias_value":"57VKYP4V","created_at":"2026-07-01T01:17:40.926935+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.26535","citing_title":"Recursive Flow Matching","ref_index":25,"is_internal_anchor":true},{"citing_arxiv_id":"2603.14135","citing_title":"Conditional flow matching for physics-constrained inverse problems with finite training data","ref_index":24,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX","json":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX.json","graph_json":"https://pith.science/api/pith-number/57VKYP4VZTBY343UO7DG2TWTYX/graph.json","events_json":"https://pith.science/api/pith-number/57VKYP4VZTBY343UO7DG2TWTYX/events.json","paper":"https://pith.science/paper/57VKYP4V"},"agent_actions":{"view_html":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX","download_json":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX.json","view_paper":"https://pith.science/paper/57VKYP4V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.09156&json=true","fetch_graph":"https://pith.science/api/pith-number/57VKYP4VZTBY343UO7DG2TWTYX/graph.json","fetch_events":"https://pith.science/api/pith-number/57VKYP4VZTBY343UO7DG2TWTYX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX/action/storage_attestation","attest_author":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX/action/author_attestation","sign_citation":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX/action/citation_signature","submit_replication":"https://pith.science/pith/57VKYP4VZTBY343UO7DG2TWTYX/action/replication_record"}},"created_at":"2026-07-01T01:17:40.926935+00:00","updated_at":"2026-07-01T01:17:40.926935+00:00"}