{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3JX7KDJR4H7FJVE52KPJDS7BON","short_pith_number":"pith:3JX7KDJR","schema_version":"1.0","canonical_sha256":"da6ff50d31e1fe54d49dd29e91cbe1734b9553bf19a687d2a44b2fd1002347ae","source":{"kind":"arxiv","id":"2603.01195","version":2},"attestation_state":"computed","paper":{"title":"VisNec: Measuring and Leveraging Visual Necessity for Multimodal Instruction Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Ren, Hongyi Cai, Jie Li, Kunyu Peng, Mingkang Dong, Sifan Zhou, Yuqian Fu","submitted_at":"2026-03-01T17:26:02Z","abstract_excerpt":"The effectiveness of multimodal instruction tuning depends not only on dataset scale, but critically on whether training samples genuinely require visual reasoning. However, existing instruction datasets often contain a substantial portion of visually redundant samples (solvable from text alone), as well as multimodally misaligned supervision that can degrade learning. To address this, we propose VisNec (Visual Necessity Score), a principled data selection framework that measures the marginal contribution of visual input during instruction tuning. By comparing predictive loss with and without "},"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":"2603.01195","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-03-01T17:26:02Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"1565d6a174fb684d5579ccd2b8a08bc1444e92f83973d88f71d810f9707d629b","abstract_canon_sha256":"4ba73d05128ab4dd0c9c6dc59bb4e5af942988c4da7f7ffba1a1c867738d0f62"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-26T01:15:49.934166Z","signature_b64":"D3MnmNprQ3c9JRyfujHVxByxOOxsvU72iWrN9ppAB7xU1xnxfhp1KHSB/8JwX86g8ZzEH+ilpwXNYps8cr4WBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"da6ff50d31e1fe54d49dd29e91cbe1734b9553bf19a687d2a44b2fd1002347ae","last_reissued_at":"2026-06-26T01:15:49.933705Z","signature_status":"signed_v1","first_computed_at":"2026-06-26T01:15:49.933705Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VisNec: Measuring and Leveraging Visual Necessity for Multimodal Instruction Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Bin Ren, Hongyi Cai, Jie Li, Kunyu Peng, Mingkang Dong, Sifan Zhou, Yuqian Fu","submitted_at":"2026-03-01T17:26:02Z","abstract_excerpt":"The effectiveness of multimodal instruction tuning depends not only on dataset scale, but critically on whether training samples genuinely require visual reasoning. However, existing instruction datasets often contain a substantial portion of visually redundant samples (solvable from text alone), as well as multimodally misaligned supervision that can degrade learning. To address this, we propose VisNec (Visual Necessity Score), a principled data selection framework that measures the marginal contribution of visual input during instruction tuning. By comparing predictive loss with and without "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.01195","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/2603.01195/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":"2603.01195","created_at":"2026-06-26T01:15:49.933762+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.01195v2","created_at":"2026-06-26T01:15:49.933762+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.01195","created_at":"2026-06-26T01:15:49.933762+00:00"},{"alias_kind":"pith_short_12","alias_value":"3JX7KDJR4H7F","created_at":"2026-06-26T01:15:49.933762+00:00"},{"alias_kind":"pith_short_16","alias_value":"3JX7KDJR4H7FJVE5","created_at":"2026-06-26T01:15:49.933762+00:00"},{"alias_kind":"pith_short_8","alias_value":"3JX7KDJR","created_at":"2026-06-26T01:15:49.933762+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/3JX7KDJR4H7FJVE52KPJDS7BON","json":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON.json","graph_json":"https://pith.science/api/pith-number/3JX7KDJR4H7FJVE52KPJDS7BON/graph.json","events_json":"https://pith.science/api/pith-number/3JX7KDJR4H7FJVE52KPJDS7BON/events.json","paper":"https://pith.science/paper/3JX7KDJR"},"agent_actions":{"view_html":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON","download_json":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON.json","view_paper":"https://pith.science/paper/3JX7KDJR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.01195&json=true","fetch_graph":"https://pith.science/api/pith-number/3JX7KDJR4H7FJVE52KPJDS7BON/graph.json","fetch_events":"https://pith.science/api/pith-number/3JX7KDJR4H7FJVE52KPJDS7BON/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON/action/storage_attestation","attest_author":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON/action/author_attestation","sign_citation":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON/action/citation_signature","submit_replication":"https://pith.science/pith/3JX7KDJR4H7FJVE52KPJDS7BON/action/replication_record"}},"created_at":"2026-06-26T01:15:49.933762+00:00","updated_at":"2026-06-26T01:15:49.933762+00:00"}