{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:NNEEBC6MNLTJFFST3PZ52AXYOP","short_pith_number":"pith:NNEEBC6M","schema_version":"1.0","canonical_sha256":"6b48408bcc6ae6929653dbf3dd02f873da577f53015ee083b816bbee0daab48b","source":{"kind":"arxiv","id":"2406.10995","version":2},"attestation_state":"computed","paper":{"title":"Concept-skill Transferability-based Data Selection for Large Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Boyang Li, Jaewoo Lee, Sung Ju Hwang","submitted_at":"2024-06-16T16:15:20Z","abstract_excerpt":"Instruction tuning, or supervised finetuning on extensive task-specific data, is necessary for Large Vision-Language Models (LVLMs) to generalize well across a broad range of vision-language (VL) tasks. However, training on large VL datasets can become prohibitively expensive. In this work, we introduce COINCIDE, an effective and scalable data selection technique that uses a small model as a reference model to select visual instruction tuning data for efficient finetuning of a target LVLM, focusing on diversity and transferability. Specifically, we cluster the training data using internal acti"},"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":"2406.10995","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-06-16T16:15:20Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8743576bc2ff330debc6123b95c568e775ed493f3703e488c69fca57fe2c50d4","abstract_canon_sha256":"a989c245a9987be6d0b311f124e165f47a66bff359b261fffada9fb003231e05"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:14:34.652868Z","signature_b64":"bVBb0fZDpkAAu8bLeEJ6Byf92Dd5mEUylvfXgbLWoN2VQc7Q8MkNXN7Z8XbMAbMuTUTmEwBuaoeVB1T0M/sKBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6b48408bcc6ae6929653dbf3dd02f873da577f53015ee083b816bbee0daab48b","last_reissued_at":"2026-07-05T09:14:34.652451Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:14:34.652451Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Concept-skill Transferability-based Data Selection for Large Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Boyang Li, Jaewoo Lee, Sung Ju Hwang","submitted_at":"2024-06-16T16:15:20Z","abstract_excerpt":"Instruction tuning, or supervised finetuning on extensive task-specific data, is necessary for Large Vision-Language Models (LVLMs) to generalize well across a broad range of vision-language (VL) tasks. However, training on large VL datasets can become prohibitively expensive. In this work, we introduce COINCIDE, an effective and scalable data selection technique that uses a small model as a reference model to select visual instruction tuning data for efficient finetuning of a target LVLM, focusing on diversity and transferability. Specifically, we cluster the training data using internal acti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.10995","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/2406.10995/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":"2406.10995","created_at":"2026-07-05T09:14:34.652512+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.10995v2","created_at":"2026-07-05T09:14:34.652512+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.10995","created_at":"2026-07-05T09:14:34.652512+00:00"},{"alias_kind":"pith_short_12","alias_value":"NNEEBC6MNLTJ","created_at":"2026-07-05T09:14:34.652512+00:00"},{"alias_kind":"pith_short_16","alias_value":"NNEEBC6MNLTJFFST","created_at":"2026-07-05T09:14:34.652512+00:00"},{"alias_kind":"pith_short_8","alias_value":"NNEEBC6M","created_at":"2026-07-05T09:14:34.652512+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26761","citing_title":"Once-For-All: A Train-Once and Select-Anytime Framework for Multimodal Instruction Tuning","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2504.21850","citing_title":"Visual Compositional Tuning","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP","json":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP.json","graph_json":"https://pith.science/api/pith-number/NNEEBC6MNLTJFFST3PZ52AXYOP/graph.json","events_json":"https://pith.science/api/pith-number/NNEEBC6MNLTJFFST3PZ52AXYOP/events.json","paper":"https://pith.science/paper/NNEEBC6M"},"agent_actions":{"view_html":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP","download_json":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP.json","view_paper":"https://pith.science/paper/NNEEBC6M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.10995&json=true","fetch_graph":"https://pith.science/api/pith-number/NNEEBC6MNLTJFFST3PZ52AXYOP/graph.json","fetch_events":"https://pith.science/api/pith-number/NNEEBC6MNLTJFFST3PZ52AXYOP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP/action/storage_attestation","attest_author":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP/action/author_attestation","sign_citation":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP/action/citation_signature","submit_replication":"https://pith.science/pith/NNEEBC6MNLTJFFST3PZ52AXYOP/action/replication_record"}},"created_at":"2026-07-05T09:14:34.652512+00:00","updated_at":"2026-07-05T09:14:34.652512+00:00"}