{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:BTEE4D42NLR6F3KFQYY5KZCAV3","short_pith_number":"pith:BTEE4D42","schema_version":"1.0","canonical_sha256":"0cc84e0f9a6ae3e2ed458631d56440aed7be75312e86c452e46624dd1c07b07e","source":{"kind":"arxiv","id":"2410.17637","version":1},"attestation_state":"computed","paper":{"title":"MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Conghui He, Dahua Lin, Haodong Duan, Jiaqi Wang, Pan Zhang, Xiaoyi Dong, Yuanjun Xiong, Yuhang Cao, Yuhang Zang, Ziyu Liu","submitted_at":"2024-10-23T07:56:48Z","abstract_excerpt":"Visual preference alignment involves training Large Vision-Language Models (LVLMs) to predict human preferences between visual inputs. This is typically achieved by using labeled datasets of chosen/rejected pairs and employing optimization algorithms like direct preference optimization (DPO). Existing visual alignment methods, primarily designed for single-image scenarios, struggle to effectively handle the complexity of multi-image tasks due to the scarcity of diverse training data and the high cost of annotating chosen/rejected pairs. We present Multi-Image Augmented Direct Preference Optimi"},"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":"2410.17637","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2024-10-23T07:56:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"35fcc7c5cf76159be08a2586391c1fe84106b7048a2719e4c75b5359102b5a43","abstract_canon_sha256":"e26b8e5e3ed592347b1874d7a10c798a7fe0d8e2d4c780e8b340a206e5e8f3e2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:41.967191Z","signature_b64":"gh/JxjHua/HBZ4Z/UDELRUhHeNvM7P9Xts4m5yDqu5tNmb3a0av/CFRuS3Y4FVh5x87YYThgsze+QkkbDSEzCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0cc84e0f9a6ae3e2ed458631d56440aed7be75312e86c452e46624dd1c07b07e","last_reissued_at":"2026-07-05T09:24:41.966569Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:41.966569Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MIA-DPO: Multi-Image Augmented Direct Preference Optimization For Large Vision-Language Models","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Conghui He, Dahua Lin, Haodong Duan, Jiaqi Wang, Pan Zhang, Xiaoyi Dong, Yuanjun Xiong, Yuhang Cao, Yuhang Zang, Ziyu Liu","submitted_at":"2024-10-23T07:56:48Z","abstract_excerpt":"Visual preference alignment involves training Large Vision-Language Models (LVLMs) to predict human preferences between visual inputs. This is typically achieved by using labeled datasets of chosen/rejected pairs and employing optimization algorithms like direct preference optimization (DPO). Existing visual alignment methods, primarily designed for single-image scenarios, struggle to effectively handle the complexity of multi-image tasks due to the scarcity of diverse training data and the high cost of annotating chosen/rejected pairs. We present Multi-Image Augmented Direct Preference Optimi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.17637","kind":"arxiv","version":1},"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/2410.17637/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":"2410.17637","created_at":"2026-07-05T09:24:41.966654+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.17637v1","created_at":"2026-07-05T09:24:41.966654+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.17637","created_at":"2026-07-05T09:24:41.966654+00:00"},{"alias_kind":"pith_short_12","alias_value":"BTEE4D42NLR6","created_at":"2026-07-05T09:24:41.966654+00:00"},{"alias_kind":"pith_short_16","alias_value":"BTEE4D42NLR6F3KF","created_at":"2026-07-05T09:24:41.966654+00:00"},{"alias_kind":"pith_short_8","alias_value":"BTEE4D42","created_at":"2026-07-05T09:24:41.966654+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.31933","citing_title":"No Place to Hide: Benchmarking Video Hallucination with Background-Controlled Pairs","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26396","citing_title":"Advancing Creative Physical Intelligence in Large Multimodal Models","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2503.01785","citing_title":"Visual-RFT: Visual Reinforcement Fine-Tuning","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04874","citing_title":"Uncertainty-Aware Exploratory Direct Preference Optimization for Multimodal Large Language Models","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19406","citing_title":"HP-Edit: A Human-Preference Post-Training Framework for Image Editing","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21911","citing_title":"When Prompts Override Vision: Prompt-Induced Hallucinations in LVLMs","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3","json":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3.json","graph_json":"https://pith.science/api/pith-number/BTEE4D42NLR6F3KFQYY5KZCAV3/graph.json","events_json":"https://pith.science/api/pith-number/BTEE4D42NLR6F3KFQYY5KZCAV3/events.json","paper":"https://pith.science/paper/BTEE4D42"},"agent_actions":{"view_html":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3","download_json":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3.json","view_paper":"https://pith.science/paper/BTEE4D42","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.17637&json=true","fetch_graph":"https://pith.science/api/pith-number/BTEE4D42NLR6F3KFQYY5KZCAV3/graph.json","fetch_events":"https://pith.science/api/pith-number/BTEE4D42NLR6F3KFQYY5KZCAV3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3/action/storage_attestation","attest_author":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3/action/author_attestation","sign_citation":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3/action/citation_signature","submit_replication":"https://pith.science/pith/BTEE4D42NLR6F3KFQYY5KZCAV3/action/replication_record"}},"created_at":"2026-07-05T09:24:41.966654+00:00","updated_at":"2026-07-05T09:24:41.966654+00:00"}