{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T03:32:09Z","timestamp":1777865529050,"version":"3.51.4"},"reference-count":82,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,10,19]],"date-time":"2025-10-19T00:00:00Z","timestamp":1760832000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key R&D Program of China","doi-asserted-by":"publisher","award":["2022ZD0115502"],"award-info":[{"award-number":["2022ZD0115502"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62461160308,U23B2010"],"award-info":[{"award-number":["62461160308,U23B2010"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,10,19]]},"DOI":"10.1109\/iccv51701.2025.00195","type":"proceedings-article","created":{"date-parts":[[2026,4,29]],"date-time":"2026-04-29T19:45:49Z","timestamp":1777491949000},"page":"2010-2021","source":"Crossref","is-referenced-by-count":0,"title":["Instruction-Oriented Preference Alignment for Enhancing Multi-Modal Comprehension Capability of MLLMs"],"prefix":"10.1109","author":[{"given":"Zitian","family":"Wang","sequence":"first","affiliation":[{"name":"Beihang University"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yue","family":"Liao","sequence":"additional","affiliation":[{"name":"National University of Singapore"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kang","family":"Rong","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Fengyun","family":"Rao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yibo","family":"Yang","sequence":"additional","affiliation":[{"name":"King Abdullah University of Science and Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Si","family":"Liu","sequence":"additional","affiliation":[{"name":"Beihang University"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","author":"Bai","year":"2025","journal-title":"Qwen2.5-vl technical report"},{"key":"ref2","author":"Bai","year":"2025","journal-title":"Qwen2. 5-vl technical report"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.2307\/2334029"},{"key":"ref4","author":"Hardy Chen","year":"2024","journal-title":"Allava: Harnessing gpt4v-synthesized data for a lite vision-language model"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72643-9_22"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.52202\/079017-0850"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.446"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4231-5"},{"key":"ref9","author":"Chen","year":"2025","journal-title":"Expanding performance boundaries of open-source multimodal models with model, data, and test-time scaling"},{"key":"ref10","volume-title":"Opencompass: A universal evaluation platform for foundation models","year":"2023"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.00018"},{"key":"ref12","author":"Deng","year":"2024","journal-title":"Efficient self-improvement in multimodal large language models: A model-level judge-free approach"},{"key":"ref13","author":"Dong","year":"2023","journal-title":"Raft: Reward ranked finetuning for generative foundation model alignment"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3664647.3685520"},{"key":"ref15","author":"Fu","year":"2024","journal-title":"Mme: A comprehensive evaluation benchmark for multimodal large language models"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.670"},{"key":"ref17","author":"Gu","year":"2025","journal-title":"Infinity-mm: Scaling multimodal performance with largescale and high-quality instruction data"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01363"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v38i16.29771"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.acl-long.680"},{"key":"ref21","author":"He","year":"2024","journal-title":"A topic-level self-correctional approach to mitigate hallucinations in mllms"},{"key":"ref22","author":"He","year":"2024","journal-title":"A topic-level self-correctional approach to mitigate hallucinations in mllms"},{"issue":"2","key":"ref23","article-title":"Lora: Low-rank adaptation of large language models","volume":"1","author":"Hu","year":"2022","journal-title":"ICLR"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00686"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00592"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.704"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52733.2024.01316"},{"key":"ref28","author":"Li","year":"2024","journal-title":"Llava-onevision: Easy visual task transfer"},{"key":"ref29","author":"Li","year":"2024","journal-title":"Llava-next-interleave: Tackling multi-image, video, and 3d in large multimodal models"},{"key":"ref30","author":"Li","year":"2023","journal-title":"Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models"},{"key":"ref31","author":"Li","year":"2023","journal-title":"Silkie: Preference distillation for large visual language models"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.358"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.358"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.20"},{"key":"ref35","author":"Li","year":"2025","journal-title":"Eagle 2: Building post-training data strategies from scratch for frontier vision-language models"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.nlp4convai-1.5"},{"key":"ref37","author":"Lindstr\u00f6m","year":"2022","journal-title":"Clevr-math: A dataset for compositional language, visual and mathematical reasoning"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1580"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1516"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02484"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1007\/s11432-024-4235-6"},{"key":"ref42","author":"Liu","year":"2024","journal-title":"Mia-dpo: Multi-image augmented direct preference optimization for large vision-language models"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.52202\/068431-0182"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.findings-acl.1009"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00331"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.findings-acl.177"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/WACV48630.2021.00225"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/WACV51458.2022.00264"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73397-0_18"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73116-7_23"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.52202\/068431-2011"},{"key":"ref52","author":"Peng","year":"2023","journal-title":"Kosmos-2: Grounding multimodal large language models to the world"},{"key":"ref53","author":"Putta","year":"2024","journal-title":"Agent q: Advanced reasoning and learning for autonomous ai agents"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2338"},{"key":"ref55","author":"Schulman","year":"2017","journal-title":"Proximal policy optimization algorithms"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20074-8_9"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0377"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58536-5_44"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00851"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.775"},{"key":"ref61","author":"Team","year":"2023","journal-title":"Millican, Katie. Gemini: a family of highly capable multimodal models"},{"key":"ref62","author":"Wang","year":"2024","journal-title":"Halu-j: Critique-based hallucination judge"},{"key":"ref63","author":"Wang","year":"2024","journal-title":"Enhancing the reasoning ability of multimodal large language models via mixed preference optimization"},{"key":"ref64","author":"Wang","year":"2024","journal-title":"Speculative rag: Enhancing retrieval augmented generation through drafting"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01271"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01271"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-acl.905"},{"key":"ref68","author":"Yang","year":"2024","journal-title":"Qwen2. 5-math technical report: Toward mathematical expert model via self-improvement"},{"key":"ref69","author":"Yao","year":"2024","journal-title":"Minicpm-v: A gpt-4v level mllm on your phone"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01230"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.01310"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/cvpr52734.2025.01861"},{"key":"ref73","author":"Yu","year":"2024","journal-title":"Rlaif-v: Aligning mllms through open-source ai feedback for super gpt-4v trustworthiness"},{"key":"ref74","author":"Yu","year":"2024","journal-title":"Mm -vet: Evaluating large multimodal models for integrated capabilities"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.52202\/075280-0482"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00913"},{"key":"ref77","author":"Yuksekgonul","year":"2023","journal-title":"When and why visionlanguage models behave like bags-of-words, and what to do about it?"},{"key":"ref78","author":"Zhang","year":"2024","journal-title":"Llavar: Enhanced visual instruction tuning for text-rich image understanding"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/icme59968.2025.11209377"},{"key":"ref80","author":"Zheng","year":"2023","journal-title":"Judging llm-as-a-judge with mt-bench and chatbot arena"},{"key":"ref81","author":"Zhou","year":"2024","journal-title":"Aligning modalities in vision large language models via preference fine-tuning"},{"key":"ref82","author":"Zhu","year":"2025","journal-title":"Judgelm: Fine-tuned large language models are scalable judges"}],"event":{"name":"2025 IEEE\/CVF International Conference on Computer Vision (ICCV)","location":"Honolulu, HI, USA","start":{"date-parts":[[2025,10,19]]},"end":{"date-parts":[[2025,10,25]]}},"container-title":["2025 IEEE\/CVF International Conference on Computer Vision (ICCV)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11443115\/11443287\/11444022.pdf?arnumber=11444022","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T06:46:01Z","timestamp":1777531561000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11444022\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,10,19]]},"references-count":82,"URL":"https:\/\/doi.org\/10.1109\/iccv51701.2025.00195","relation":{},"subject":[],"published":{"date-parts":[[2025,10,19]]}}}