{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:KVWEU6I4HYS4MX74JSNBYIDSDA","short_pith_number":"pith:KVWEU6I4","schema_version":"1.0","canonical_sha256":"556c4a791c3e25c65ffc4c9a1c2072180cd5fca60a8ea58f4b04c9ac64685c9a","source":{"kind":"arxiv","id":"2212.08158","version":3},"attestation_state":"computed","paper":{"title":"MM-SHAP: A Performance-agnostic Metric for Measuring Multimodal Contributions in Vision and Language Models & Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Anette Frank, Letitia Parcalabescu","submitted_at":"2022-12-15T21:41:06Z","abstract_excerpt":"Vision and language models (VL) are known to exploit unrobust indicators in individual modalities (e.g., introduced by distributional biases) instead of focusing on relevant information in each modality. That a unimodal model achieves similar accuracy on a VL task to a multimodal one, indicates that so-called unimodal collapse occurred. However, accuracy-based tests fail to detect e.g., when the model prediction is wrong, while the model used relevant information from a modality. Instead, we propose MM-SHAP, a performance-agnostic multimodality score based on Shapley values that reliably quant"},"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":"2212.08158","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-12-15T21:41:06Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"1439523760a52a8e8a8881024e6df9804470e2816248a26a135d194ce57f5d34","abstract_canon_sha256":"faa921639ebe4d988ea46bd43b1f129a52aa025f7b688bd39113c365983abd7b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:08:44.761251Z","signature_b64":"E4ObbdefV25qSCukiDFHADIeTESwZBKnssYNJ2eL8i0xR/GG5djINLd2LEhiQzJVoUIHB10B3+1ecJijuP20Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"556c4a791c3e25c65ffc4c9a1c2072180cd5fca60a8ea58f4b04c9ac64685c9a","last_reissued_at":"2026-07-05T09:08:44.760720Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:08:44.760720Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MM-SHAP: A Performance-agnostic Metric for Measuring Multimodal Contributions in Vision and Language Models & Tasks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Anette Frank, Letitia Parcalabescu","submitted_at":"2022-12-15T21:41:06Z","abstract_excerpt":"Vision and language models (VL) are known to exploit unrobust indicators in individual modalities (e.g., introduced by distributional biases) instead of focusing on relevant information in each modality. That a unimodal model achieves similar accuracy on a VL task to a multimodal one, indicates that so-called unimodal collapse occurred. However, accuracy-based tests fail to detect e.g., when the model prediction is wrong, while the model used relevant information from a modality. Instead, we propose MM-SHAP, a performance-agnostic multimodality score based on Shapley values that reliably quant"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2212.08158","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/2212.08158/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":"2212.08158","created_at":"2026-07-05T09:08:44.760787+00:00"},{"alias_kind":"arxiv_version","alias_value":"2212.08158v3","created_at":"2026-07-05T09:08:44.760787+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2212.08158","created_at":"2026-07-05T09:08:44.760787+00:00"},{"alias_kind":"pith_short_12","alias_value":"KVWEU6I4HYS4","created_at":"2026-07-05T09:08:44.760787+00:00"},{"alias_kind":"pith_short_16","alias_value":"KVWEU6I4HYS4MX74","created_at":"2026-07-05T09:08:44.760787+00:00"},{"alias_kind":"pith_short_8","alias_value":"KVWEU6I4","created_at":"2026-07-05T09:08:44.760787+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2508.04427","citing_title":"Decoding the Multimodal Maze: A Systematic Review on the Adoption of Explainability in Multimodal Attention-based Models","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA","json":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA.json","graph_json":"https://pith.science/api/pith-number/KVWEU6I4HYS4MX74JSNBYIDSDA/graph.json","events_json":"https://pith.science/api/pith-number/KVWEU6I4HYS4MX74JSNBYIDSDA/events.json","paper":"https://pith.science/paper/KVWEU6I4"},"agent_actions":{"view_html":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA","download_json":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA.json","view_paper":"https://pith.science/paper/KVWEU6I4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2212.08158&json=true","fetch_graph":"https://pith.science/api/pith-number/KVWEU6I4HYS4MX74JSNBYIDSDA/graph.json","fetch_events":"https://pith.science/api/pith-number/KVWEU6I4HYS4MX74JSNBYIDSDA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA/action/storage_attestation","attest_author":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA/action/author_attestation","sign_citation":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA/action/citation_signature","submit_replication":"https://pith.science/pith/KVWEU6I4HYS4MX74JSNBYIDSDA/action/replication_record"}},"created_at":"2026-07-05T09:08:44.760787+00:00","updated_at":"2026-07-05T09:08:44.760787+00:00"}