{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BESSIXXJ3VTMF7U6YJB3L4QOTR","short_pith_number":"pith:BESSIXXJ","schema_version":"1.0","canonical_sha256":"0925245ee9dd66c2fe9ec243b5f20e9c7829c9c3937ea1d2dd5222fc4a6848c9","source":{"kind":"arxiv","id":"2608.01462","version":1},"attestation_state":"computed","paper":{"title":"Loud or Silent? A Reusable Framework for Per-Modality Failure Analysis in Multimodal Clinical AI","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ahram Han, Aly Dhedhi, Boya Zhang, Dimitrios Proios, Jacques Kpodonu, Kacper Dobek, Kaushik Madapati, Kevin Zhou, Krittaphas Chaisutyakorn, Kushul Reddy Palakala, Leo Anthony Celi, Noah Dane Hebdon, Quang Bui, Rahul Gorijavolu, Samuel Paik-Heintz, Sebasti\\'an Andr\\'es Cajas Ord\\'o\\~nez, Shlok Jaiswal","submitted_at":"2026-08-02T19:38:56Z","abstract_excerpt":"Multimodal clinical models are usually judged on accuracy with every modality present, but deployment removes modalities; an echocardiogram is often unavailable where an ECG is routine. Two questions then matter beyond the size of the accuracy loss: which modality was responsible, and whether the model fails loudly or silently once that modality is dropped. The distinction is per-example and modality-level, and is separate from post-hoc feature attribution (e.g. SHAP). Models are replaced often; the evaluation that answers these questions is reused. We present a model-agnostic modality-failure"},"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":"2608.01462","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-02T19:38:56Z","cross_cats_sorted":[],"title_canon_sha256":"2cddf27f7f8d2f961f7b36b491b4d9aa1d9931ef7f0dd76cdfdfc34d492a95b8","abstract_canon_sha256":"b7626738a82081746a023b52996bbb5d8ff9981d29fb022adc280bc96854bf65"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:05:17.003568Z","signature_b64":"sWBG8U4fNBOU1wkXJArZTLlLS1IEAljhYXkvaBTgtV6fjp1KG61H0u2CJ/ytobymxQMxCwLFX2hZzfwLOQpcDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0925245ee9dd66c2fe9ec243b5f20e9c7829c9c3937ea1d2dd5222fc4a6848c9","last_reissued_at":"2026-08-04T02:05:17.001784Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:05:17.001784Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Loud or Silent? A Reusable Framework for Per-Modality Failure Analysis in Multimodal Clinical AI","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Ahram Han, Aly Dhedhi, Boya Zhang, Dimitrios Proios, Jacques Kpodonu, Kacper Dobek, Kaushik Madapati, Kevin Zhou, Krittaphas Chaisutyakorn, Kushul Reddy Palakala, Leo Anthony Celi, Noah Dane Hebdon, Quang Bui, Rahul Gorijavolu, Samuel Paik-Heintz, Sebasti\\'an Andr\\'es Cajas Ord\\'o\\~nez, Shlok Jaiswal","submitted_at":"2026-08-02T19:38:56Z","abstract_excerpt":"Multimodal clinical models are usually judged on accuracy with every modality present, but deployment removes modalities; an echocardiogram is often unavailable where an ECG is routine. Two questions then matter beyond the size of the accuracy loss: which modality was responsible, and whether the model fails loudly or silently once that modality is dropped. The distinction is per-example and modality-level, and is separate from post-hoc feature attribution (e.g. SHAP). Models are replaced often; the evaluation that answers these questions is reused. We present a model-agnostic modality-failure"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01462","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/2608.01462/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":"2608.01462","created_at":"2026-08-04T02:05:17.003361+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.01462v1","created_at":"2026-08-04T02:05:17.003361+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01462","created_at":"2026-08-04T02:05:17.003361+00:00"},{"alias_kind":"pith_short_12","alias_value":"BESSIXXJ3VTM","created_at":"2026-08-04T02:05:17.003361+00:00"},{"alias_kind":"pith_short_16","alias_value":"BESSIXXJ3VTMF7U6","created_at":"2026-08-04T02:05:17.003361+00:00"},{"alias_kind":"pith_short_8","alias_value":"BESSIXXJ","created_at":"2026-08-04T02:05:17.003361+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/BESSIXXJ3VTMF7U6YJB3L4QOTR","json":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR.json","graph_json":"https://pith.science/api/pith-number/BESSIXXJ3VTMF7U6YJB3L4QOTR/graph.json","events_json":"https://pith.science/api/pith-number/BESSIXXJ3VTMF7U6YJB3L4QOTR/events.json","paper":"https://pith.science/paper/BESSIXXJ"},"agent_actions":{"view_html":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR","download_json":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR.json","view_paper":"https://pith.science/paper/BESSIXXJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.01462&json=true","fetch_graph":"https://pith.science/api/pith-number/BESSIXXJ3VTMF7U6YJB3L4QOTR/graph.json","fetch_events":"https://pith.science/api/pith-number/BESSIXXJ3VTMF7U6YJB3L4QOTR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR/action/storage_attestation","attest_author":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR/action/author_attestation","sign_citation":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR/action/citation_signature","submit_replication":"https://pith.science/pith/BESSIXXJ3VTMF7U6YJB3L4QOTR/action/replication_record"}},"created_at":"2026-08-04T02:05:17.003361+00:00","updated_at":"2026-08-04T02:05:17.003361+00:00"}