{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:ZXYIHMW7N6LJ7422FEWDUZAO2I","short_pith_number":"pith:ZXYIHMW7","schema_version":"1.0","canonical_sha256":"cdf083b2df6f969ff35a292c3a640ed20f0c4e929a0c56f6324941a78d06bf6f","source":{"kind":"arxiv","id":"2412.19684","version":2},"attestation_state":"computed","paper":{"title":"Boosting Private Domain Understanding of Efficient MLLMs: A Tuning-free, Adaptive, Universal Prompt Optimization Framework","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bolin Li, Fangxun Shu, Hao Cheng, Hao Jiang, Haoyuan Li, Jiang Liu, Jinghao Wei, Juncheng Li, Siliang Tang, Tao Zhong, Tianwei Lin, Wanggui He, Wenqiao Zhang, Yueting Zhuang, Zhelun Yu, Zheqi Lv","submitted_at":"2024-12-27T15:21:17Z","abstract_excerpt":"Efficient multimodal large language models (EMLLMs), in contrast to multimodal large language models (MLLMs), reduce model size and computational costs and are often deployed on resource-constrained devices. However, due to data privacy concerns, existing open-source EMLLMs rarely have access to private domain-specific data during the pre-training process, making them difficult to directly apply in device-specific domains, such as certain business scenarios. To address this weakness, this paper focuses on the efficient adaptation of EMLLMs to private domains, specifically in two areas: 1) how "},"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":"2412.19684","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-12-27T15:21:17Z","cross_cats_sorted":[],"title_canon_sha256":"7a210acd3481064e84a485dce051bf7ea936d69eef0db527dfd45aad63541757","abstract_canon_sha256":"2ea30b8c364a8b699898c5b88cdcf3637951446ab8b79191eaf30f1cc5189515"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:24.159996Z","signature_b64":"94kxXuXKm/K/OwcdOg3oNaeNn8IWrm1ixQWKG4d6SAgiD7OkytKvLBYyOO9R/OLqLWNjG8teHw6JtOdJC3kSAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cdf083b2df6f969ff35a292c3a640ed20f0c4e929a0c56f6324941a78d06bf6f","last_reissued_at":"2026-07-05T10:15:24.159500Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:24.159500Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Boosting Private Domain Understanding of Efficient MLLMs: A Tuning-free, Adaptive, Universal Prompt Optimization Framework","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Bolin Li, Fangxun Shu, Hao Cheng, Hao Jiang, Haoyuan Li, Jiang Liu, Jinghao Wei, Juncheng Li, Siliang Tang, Tao Zhong, Tianwei Lin, Wanggui He, Wenqiao Zhang, Yueting Zhuang, Zhelun Yu, Zheqi Lv","submitted_at":"2024-12-27T15:21:17Z","abstract_excerpt":"Efficient multimodal large language models (EMLLMs), in contrast to multimodal large language models (MLLMs), reduce model size and computational costs and are often deployed on resource-constrained devices. However, due to data privacy concerns, existing open-source EMLLMs rarely have access to private domain-specific data during the pre-training process, making them difficult to directly apply in device-specific domains, such as certain business scenarios. To address this weakness, this paper focuses on the efficient adaptation of EMLLMs to private domains, specifically in two areas: 1) how "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.19684","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/2412.19684/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":"2412.19684","created_at":"2026-07-05T10:15:24.159561+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.19684v2","created_at":"2026-07-05T10:15:24.159561+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.19684","created_at":"2026-07-05T10:15:24.159561+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZXYIHMW7N6LJ","created_at":"2026-07-05T10:15:24.159561+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZXYIHMW7N6LJ7422","created_at":"2026-07-05T10:15:24.159561+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZXYIHMW7","created_at":"2026-07-05T10:15:24.159561+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2412.04300","citing_title":"T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I","json":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I.json","graph_json":"https://pith.science/api/pith-number/ZXYIHMW7N6LJ7422FEWDUZAO2I/graph.json","events_json":"https://pith.science/api/pith-number/ZXYIHMW7N6LJ7422FEWDUZAO2I/events.json","paper":"https://pith.science/paper/ZXYIHMW7"},"agent_actions":{"view_html":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I","download_json":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I.json","view_paper":"https://pith.science/paper/ZXYIHMW7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.19684&json=true","fetch_graph":"https://pith.science/api/pith-number/ZXYIHMW7N6LJ7422FEWDUZAO2I/graph.json","fetch_events":"https://pith.science/api/pith-number/ZXYIHMW7N6LJ7422FEWDUZAO2I/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I/action/storage_attestation","attest_author":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I/action/author_attestation","sign_citation":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I/action/citation_signature","submit_replication":"https://pith.science/pith/ZXYIHMW7N6LJ7422FEWDUZAO2I/action/replication_record"}},"created_at":"2026-07-05T10:15:24.159561+00:00","updated_at":"2026-07-05T10:15:24.159561+00:00"}