{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:QD76ISTWUXCHJSFJMSMRIIUR4O","short_pith_number":"pith:QD76ISTW","schema_version":"1.0","canonical_sha256":"80ffe44a76a5c474c8a96499142291e3960a97b277dfc59511518b2555243fa3","source":{"kind":"arxiv","id":"2411.19930","version":4},"attestation_state":"computed","paper":{"title":"On Domain-Adaptive Post-Training for Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bo Dai, Daixuan Cheng, Shaohan Huang, Wayne Xin Zhao, Xintong Zhang, Zhenliang Zhang, Zhongzhi Luan, Ziyu Zhu","submitted_at":"2024-11-29T18:42:28Z","abstract_excerpt":"Adapting general multimodal large language models (MLLMs) to specific domains, such as scientific and industrial fields, is highly significant in promoting their practical applications. This paper systematically investigates domain adaptation of MLLMs via post-training, focusing on data synthesis, training pipeline, and task evaluation. (1) Data Synthesis: Using only open-source models, we develop a generate-then-filter pipeline that curates diverse visual instruction tasks based on domain-specific image-caption pairs. The resulting data surpass the data synthesized by manual rules or strong c"},"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":"2411.19930","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2024-11-29T18:42:28Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"7eeccc8c740ae036848fcbf4a902a4d22ffa10c86adb6978badee7afc69c9216","abstract_canon_sha256":"d66524f9ccbaec6c80825335bfca5ddfe494c3577876b3db7ff08f51cd9544fe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:59:55.492283Z","signature_b64":"OVlZlNnMKnYSgP3AAQw4S0xMVsVmQUmu7qS4HsJJNxf8Z/U/B/uAVlo4p40La06bKd1FlLxxSon0v+MerAW9AQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"80ffe44a76a5c474c8a96499142291e3960a97b277dfc59511518b2555243fa3","last_reissued_at":"2026-07-05T11:59:55.491723Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:59:55.491723Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On Domain-Adaptive Post-Training for Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Bo Dai, Daixuan Cheng, Shaohan Huang, Wayne Xin Zhao, Xintong Zhang, Zhenliang Zhang, Zhongzhi Luan, Ziyu Zhu","submitted_at":"2024-11-29T18:42:28Z","abstract_excerpt":"Adapting general multimodal large language models (MLLMs) to specific domains, such as scientific and industrial fields, is highly significant in promoting their practical applications. This paper systematically investigates domain adaptation of MLLMs via post-training, focusing on data synthesis, training pipeline, and task evaluation. (1) Data Synthesis: Using only open-source models, we develop a generate-then-filter pipeline that curates diverse visual instruction tasks based on domain-specific image-caption pairs. The resulting data surpass the data synthesized by manual rules or strong c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.19930","kind":"arxiv","version":4},"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/2411.19930/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":"2411.19930","created_at":"2026-07-05T11:59:55.491804+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.19930v4","created_at":"2026-07-05T11:59:55.491804+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.19930","created_at":"2026-07-05T11:59:55.491804+00:00"},{"alias_kind":"pith_short_12","alias_value":"QD76ISTWUXCH","created_at":"2026-07-05T11:59:55.491804+00:00"},{"alias_kind":"pith_short_16","alias_value":"QD76ISTWUXCHJSFJ","created_at":"2026-07-05T11:59:55.491804+00:00"},{"alias_kind":"pith_short_8","alias_value":"QD76ISTW","created_at":"2026-07-05T11:59:55.491804+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.05310","citing_title":"Evaluating and Understanding Model Editing for Medical Vision Language Models","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2506.09082","citing_title":"AVA-Bench: Atomic Visual Ability Benchmark for Vision Foundation Models","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2603.13779","citing_title":"AD-Copilot: A Vision-Language Assistant for Industrial Anomaly Detection via Visual In-context Comparison","ref_index":1,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O","json":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O.json","graph_json":"https://pith.science/api/pith-number/QD76ISTWUXCHJSFJMSMRIIUR4O/graph.json","events_json":"https://pith.science/api/pith-number/QD76ISTWUXCHJSFJMSMRIIUR4O/events.json","paper":"https://pith.science/paper/QD76ISTW"},"agent_actions":{"view_html":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O","download_json":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O.json","view_paper":"https://pith.science/paper/QD76ISTW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.19930&json=true","fetch_graph":"https://pith.science/api/pith-number/QD76ISTWUXCHJSFJMSMRIIUR4O/graph.json","fetch_events":"https://pith.science/api/pith-number/QD76ISTWUXCHJSFJMSMRIIUR4O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O/action/storage_attestation","attest_author":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O/action/author_attestation","sign_citation":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O/action/citation_signature","submit_replication":"https://pith.science/pith/QD76ISTWUXCHJSFJMSMRIIUR4O/action/replication_record"}},"created_at":"2026-07-05T11:59:55.491804+00:00","updated_at":"2026-07-05T11:59:55.491804+00:00"}