{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:HEUKQMNYONJ3WPLCZCAN5DB2QI","short_pith_number":"pith:HEUKQMNY","schema_version":"1.0","canonical_sha256":"3928a831b87353bb3d62c880de8c3a823894906b730eac4cf2e768f6bd4ff3f9","source":{"kind":"arxiv","id":"2503.20502","version":2},"attestation_state":"computed","paper":{"title":"MLLM-Selector: Necessity and Diversity-driven High-Value Data Selection for Enhanced Visual Instruction Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Debing Zhang, Guohai Xu, Jiayi Ji, Jie Lou, Rongrong Ji, Xiaoshuai Sun, Yiwei Ma","submitted_at":"2025-03-26T12:42:37Z","abstract_excerpt":"Visual instruction tuning (VIT) has emerged as a crucial technique for enabling multi-modal large language models (MLLMs) to follow user instructions adeptly. Yet, a significant gap persists in understanding the attributes of high-quality instruction tuning data and frameworks for its automated selection. To address this, we introduce MLLM-Selector, an automated approach that identifies valuable data for VIT by weighing necessity and diversity. Our process starts by randomly sampling a subset from the VIT data pool to fine-tune a pretrained model, thus creating a seed model with an initial abi"},"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":"2503.20502","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-26T12:42:37Z","cross_cats_sorted":[],"title_canon_sha256":"c7e17c6066fda355f44a22a9e8a5db92231c7fdef4fad6aef89bfeb87452b462","abstract_canon_sha256":"3d623de2562222772e49c891006299d5310c23a541fd80c09c566a47e4cfb9cb"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:41:23.415159Z","signature_b64":"puznyK2+2qUfPRpGLEraRdzHtIA0/YyXF46B9lmm2pUpxcdj+YvxUzLKY3P5vFC6POQJmxsoDOL5/VLdMZRqBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3928a831b87353bb3d62c880de8c3a823894906b730eac4cf2e768f6bd4ff3f9","last_reissued_at":"2026-07-05T10:41:23.414647Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:41:23.414647Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MLLM-Selector: Necessity and Diversity-driven High-Value Data Selection for Enhanced Visual Instruction Tuning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Debing Zhang, Guohai Xu, Jiayi Ji, Jie Lou, Rongrong Ji, Xiaoshuai Sun, Yiwei Ma","submitted_at":"2025-03-26T12:42:37Z","abstract_excerpt":"Visual instruction tuning (VIT) has emerged as a crucial technique for enabling multi-modal large language models (MLLMs) to follow user instructions adeptly. Yet, a significant gap persists in understanding the attributes of high-quality instruction tuning data and frameworks for its automated selection. To address this, we introduce MLLM-Selector, an automated approach that identifies valuable data for VIT by weighing necessity and diversity. Our process starts by randomly sampling a subset from the VIT data pool to fine-tune a pretrained model, thus creating a seed model with an initial abi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.20502","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/2503.20502/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":"2503.20502","created_at":"2026-07-05T10:41:23.414709+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.20502v2","created_at":"2026-07-05T10:41:23.414709+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.20502","created_at":"2026-07-05T10:41:23.414709+00:00"},{"alias_kind":"pith_short_12","alias_value":"HEUKQMNYONJ3","created_at":"2026-07-05T10:41:23.414709+00:00"},{"alias_kind":"pith_short_16","alias_value":"HEUKQMNYONJ3WPLC","created_at":"2026-07-05T10:41:23.414709+00:00"},{"alias_kind":"pith_short_8","alias_value":"HEUKQMNY","created_at":"2026-07-05T10:41:23.414709+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10651","citing_title":"Kwai Keye-VL-2.0 Technical Report","ref_index":65,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI","json":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI.json","graph_json":"https://pith.science/api/pith-number/HEUKQMNYONJ3WPLCZCAN5DB2QI/graph.json","events_json":"https://pith.science/api/pith-number/HEUKQMNYONJ3WPLCZCAN5DB2QI/events.json","paper":"https://pith.science/paper/HEUKQMNY"},"agent_actions":{"view_html":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI","download_json":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI.json","view_paper":"https://pith.science/paper/HEUKQMNY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.20502&json=true","fetch_graph":"https://pith.science/api/pith-number/HEUKQMNYONJ3WPLCZCAN5DB2QI/graph.json","fetch_events":"https://pith.science/api/pith-number/HEUKQMNYONJ3WPLCZCAN5DB2QI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI/action/storage_attestation","attest_author":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI/action/author_attestation","sign_citation":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI/action/citation_signature","submit_replication":"https://pith.science/pith/HEUKQMNYONJ3WPLCZCAN5DB2QI/action/replication_record"}},"created_at":"2026-07-05T10:41:23.414709+00:00","updated_at":"2026-07-05T10:41:23.414709+00:00"}