{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4PRMUGOXSYRWGR3POM7UAUBIO2","short_pith_number":"pith:4PRMUGOX","schema_version":"1.0","canonical_sha256":"e3e2ca19d7962363476f733f40502876b57593779c266683967926c9054589b8","source":{"kind":"arxiv","id":"2605.15235","version":1},"attestation_state":"computed","paper":{"title":"MuteBench: Modality Unavailability Tolerance Evaluation for Incomplete Multimodal Fusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Architecture family predicts robustness to missing modalities better than model size in clinical fusion.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chen Chen, Song Wang, Tianlong Chen, Wugeng Zheng, Ziwen Kan","submitted_at":"2026-05-13T19:32:11Z","abstract_excerpt":"Multimodal physiological data powers clinical AI systems from intensive care units to wearable devices, but sensors routinely fail in practice. Two failure modes are common: modality missing, where an entire channel is absent, and within-modality missing, where a contiguous time segment is lost. No existing benchmark evaluates multiple fusion architectures under both failure modes at controlled severity levels across diverse clinical datasets. We present MuteBench, a benchmark covering 9 datasets from 7 clinical domains, 6 fusion architectures, and 2 missing-data modes over 125,000 samples. Th"},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.15235","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T19:32:11Z","cross_cats_sorted":[],"title_canon_sha256":"73214917211d846889c2eedb50f229c1143ce1e33b627ce0ef3f22fda11508fc","abstract_canon_sha256":"63a8801f3385b7f89066309346fc4efe2d69bd7fa8c5d3ff52ecdeeae840a831"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:47.720758Z","signature_b64":"NtkpG8226w45WuiVnJ5Xd9miV43+byN8hWwmM15L66TQ0C6EvqaY8VPj4+PAOZtR9+/7UYdBb6rctISwotb6CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e3e2ca19d7962363476f733f40502876b57593779c266683967926c9054589b8","last_reissued_at":"2026-05-20T00:00:47.719823Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:47.719823Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MuteBench: Modality Unavailability Tolerance Evaluation for Incomplete Multimodal Fusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Architecture family predicts robustness to missing modalities better than model size in clinical fusion.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chen Chen, Song Wang, Tianlong Chen, Wugeng Zheng, Ziwen Kan","submitted_at":"2026-05-13T19:32:11Z","abstract_excerpt":"Multimodal physiological data powers clinical AI systems from intensive care units to wearable devices, but sensors routinely fail in practice. Two failure modes are common: modality missing, where an entire channel is absent, and within-modality missing, where a contiguous time segment is lost. No existing benchmark evaluates multiple fusion architectures under both failure modes at controlled severity levels across diverse clinical datasets. We present MuteBench, a benchmark covering 9 datasets from 7 clinical domains, 6 fusion architectures, and 2 missing-data modes over 125,000 samples. Th"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"architecture family is the strongest predictor of robustness, outweighing parameter count","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The nine chosen clinical datasets and six fusion architectures sufficiently represent the range of real-world multimodal physiological signals and sensor-failure patterns so that the observed robustness rankings generalize beyond the tested cases. (Abstract)","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MuteBench evaluates multimodal fusion robustness to modality missing and within-modality missing on 125000 samples from 9 clinical datasets, finding architecture family predicts tolerance better than parameter count.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Architecture family predicts robustness to missing modalities better than model size in clinical fusion.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"423ce505b8717e52ff4103d1cebd7ab21cf6c4474081c58f53e9f98b17a5c526"},"source":{"id":"2605.15235","kind":"arxiv","version":1},"verdict":{"id":"f17af8ad-701b-4dd4-a586-b674601037af","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T16:29:45.622394Z","strongest_claim":"architecture family is the strongest predictor of robustness, outweighing parameter count","one_line_summary":"MuteBench evaluates multimodal fusion robustness to modality missing and within-modality missing on 125000 samples from 9 clinical datasets, finding architecture family predicts tolerance better than parameter count.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The nine chosen clinical datasets and six fusion architectures sufficiently represent the range of real-world multimodal physiological signals and sensor-failure patterns so that the observed robustness rankings generalize beyond the tested cases. (Abstract)","pith_extraction_headline":"Architecture family predicts robustness to missing modalities better than model size in clinical fusion."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15235/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-19T17:41:56.139484Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T17:01:18.468993Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T16:36:43.039369Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.826836Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"86864cf054fc1f6e7ae938b499a98e38af15e90dd58d53609545cf7b305efaa4"},"references":{"count":63,"sample":[{"doi":"10.1038/s41598-026-39035-z","year":2026,"title":"doi: 10.1038/s41598-026-39035-z","work_id":"9837de1e-f997-48bf-941b-e515f83283c0","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1038/s41591-022-01981-2","year":2022,"title":"Multimodal biomedical AI.Nature Medicine, 28(9):1773–1784, 2022","work_id":"8658170a-8b8a-4db9-877f-2021053de5e4","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Ehrxqa: A multi-modal ques- tion answering dataset for electronic health records with chest x-ray images","work_id":"859bc897-b6db-4296-a896-c29001d8e3d2","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1145/1553374.1553380","year":2009,"title":"In Proceedings of the 26th Annual International Conference on Machine Learning (Montreal, Quebec, Canada) (ICML ’09)","work_id":"5d30e737-af53-4f1d-8f90-c8e0159a22ba","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Recurrent neural networks for multivariate time series with missing values.Scientific Reports, 8(1):6085,","work_id":"c7da1c42-db7d-409c-b51a-f2a527136b5e","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":63,"snapshot_sha256":"1e19b346919c5010097f620cf0de7911d5d0e3c42c4b08281722e1a9176dc765","internal_anchors":1},"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":"2605.15235","created_at":"2026-05-20T00:00:47.719954+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15235v1","created_at":"2026-05-20T00:00:47.719954+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15235","created_at":"2026-05-20T00:00:47.719954+00:00"},{"alias_kind":"pith_short_12","alias_value":"4PRMUGOXSYRW","created_at":"2026-05-20T00:00:47.719954+00:00"},{"alias_kind":"pith_short_16","alias_value":"4PRMUGOXSYRWGR3P","created_at":"2026-05-20T00:00:47.719954+00:00"},{"alias_kind":"pith_short_8","alias_value":"4PRMUGOX","created_at":"2026-05-20T00:00:47.719954+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/4PRMUGOXSYRWGR3POM7UAUBIO2","json":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2.json","graph_json":"https://pith.science/api/pith-number/4PRMUGOXSYRWGR3POM7UAUBIO2/graph.json","events_json":"https://pith.science/api/pith-number/4PRMUGOXSYRWGR3POM7UAUBIO2/events.json","paper":"https://pith.science/paper/4PRMUGOX"},"agent_actions":{"view_html":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2","download_json":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2.json","view_paper":"https://pith.science/paper/4PRMUGOX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15235&json=true","fetch_graph":"https://pith.science/api/pith-number/4PRMUGOXSYRWGR3POM7UAUBIO2/graph.json","fetch_events":"https://pith.science/api/pith-number/4PRMUGOXSYRWGR3POM7UAUBIO2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2/action/storage_attestation","attest_author":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2/action/author_attestation","sign_citation":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2/action/citation_signature","submit_replication":"https://pith.science/pith/4PRMUGOXSYRWGR3POM7UAUBIO2/action/replication_record"}},"created_at":"2026-05-20T00:00:47.719954+00:00","updated_at":"2026-05-20T00:00:47.719954+00:00"}