{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,16]],"date-time":"2026-06-16T15:09:29Z","timestamp":1781622569782,"version":"3.54.5"},"reference-count":81,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62136005"],"award-info":[{"award-number":["62136005"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100017607","name":"Shenzhen Municipal Fundamental Research Program","doi-asserted-by":"publisher","award":["JCYJ20250604144724032"],"award-info":[{"award-number":["JCYJ20250604144724032"]}],"id":[{"id":"10.13039\/501100017607","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Research Grants Council of Hong Kong Special Administrative Region, China","award":["16202523"],"award-info":[{"award-number":["16202523"]}]},{"name":"Research Grants Council of Hong Kong Special Administrative Region, China","award":["C7004-22G-1"],"award-info":[{"award-number":["C7004-22G-1"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. on Image Process."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/tip.2026.3678087","type":"journal-article","created":{"date-parts":[[2026,4,13]],"date-time":"2026-04-13T19:38:52Z","timestamp":1776109132000},"page":"3881-3892","source":"Crossref","is-referenced-by-count":5,"title":["Mixture of Cluster-Conditional LoRA Experts for Vision-Language Instruction Tuning"],"prefix":"10.1109","volume":"35","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-0888-4064","authenticated-orcid":false,"given":"Yunhao","family":"Gou","sequence":"first","affiliation":[{"name":"Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhili","family":"Liu","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Sai Kung, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Kai","family":"Chen","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Sai Kung, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2752-5942","authenticated-orcid":false,"given":"Lanqing","family":"Hong","sequence":"additional","affiliation":[{"name":"Huawei Noah’s Ark Laboratory, Hong Kong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hang","family":"Xu","sequence":"additional","affiliation":[{"name":"Huawei Noah’s Ark Laboratory, Hong Kong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8492-3069","authenticated-orcid":false,"given":"Zhenguo","family":"Li","sequence":"additional","affiliation":[{"name":"Huawei Noah’s Ark Laboratory, Hong Kong, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3716-8125","authenticated-orcid":false,"given":"Dit-Yan","family":"Yeung","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Sai Kung, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4828-8248","authenticated-orcid":false,"given":"James T.","family":"Kwok","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Sai Kung, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-1100-4835","authenticated-orcid":false,"given":"Yu","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Computer Science and Engineering, Southern University of Science and Technology, Shenzhen, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Proc. NeurIPS","author":"Brown"},{"key":"ref2","volume-title":"Introducing Chatgpt","year":"2022"},{"key":"ref3","article-title":"Gaining wisdom from setbacks: Aligning large language models via mistake analysis","author":"Chen","year":"2023","journal-title":"arXiv:2310.10477"},{"key":"ref4","article-title":"The flan collection: Designing data and methods for effective instruction tuning","author":"Longpre","year":"2023","journal-title":"arXiv:2301.13688"},{"key":"ref5","article-title":"Qwen-VL: A frontier large vision-language model with versatile abilities","author":"Bai","year":"2023","journal-title":"arXiv:2308.12966"},{"key":"ref6","article-title":"InternLM-XComposer: A vision-language large model for advanced text-image comprehension and composition","author":"Zhang","year":"2023","journal-title":"arXiv:2309.15112"},{"key":"ref7","article-title":"mPLUG-Owl2: Revolutionizing multi-modal large language model with modality collaboration","author":"Ye","year":"2023","journal-title":"arXiv:2311.04257"},{"key":"ref8","article-title":"ShareGPT4 V: Improving large multi-modal models with better captions","author":"Chen","year":"2023","journal-title":"arXiv:2311.12793"},{"key":"ref9","article-title":"MiniGPT-v2: Large language model as a unified interface for vision-language multi-task learning","author":"Chen","year":"2023","journal-title":"arXiv:2310.09478"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.52202\/075280-1516"},{"key":"ref11","article-title":"Improved baselines with visual instruction tuning","author":"Liu","year":"2023","journal-title":"arXiv:2310.03744"},{"key":"ref12","article-title":"MiniGPT-4: Enhancing vision-language understanding with advanced large language models","author":"Zhu","year":"2023","journal-title":"arXiv:2304.10592"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2142"},{"key":"ref14","article-title":"Multitask prompted training enables zero-shot task generalization","author":"Sanh","year":"2021","journal-title":"arXiv:2110.08207"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.emnlp-main.340"},{"key":"ref16","article-title":"Scaling instruction-finetuned language models","author":"Won Chung","year":"2022","journal-title":"arXiv:2210.11416"},{"key":"ref17","article-title":"SVIT: Scaling up visual instruction tuning","author":"Zhao","year":"2023","journal-title":"arXiv:2307.04087"},{"key":"ref18","article-title":"M3IT: A large-scale dataset towards multi-modal multilingual instruction tuning","author":"Li","year":"2023","journal-title":"arXiv:2306.04387"},{"key":"ref19","article-title":"Finetuned language models are zero-shot learners","author":"Wei","year":"2021","journal-title":"arXiv:2109.01652"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2021.3070203"},{"key":"ref21","article-title":"Deep safe multi-task learning","author":"Yue","year":"2021","journal-title":"arXiv:2111.10601"},{"key":"ref22","article-title":"LoRA: Low-rank adaptation of large language models","author":"Hu","year":"2021","journal-title":"arXiv:2106.09685"},{"key":"ref23","article-title":"Octavius: Mitigating task interference in MLLMs via LoRA-MoE","author":"Chen","year":"2023","journal-title":"arXiv:2311.02684"},{"key":"ref24","article-title":"LLaVA-MoLE: Sparse mixture of LoRA experts for mitigating data conflicts in instruction finetuning MLLMs","author":"Chen","year":"2024","journal-title":"arXiv:2401.16160"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v36i2.20079"},{"key":"ref26","article-title":"Task-customized masked AutoEncoder via mixture of cluster-conditional experts","volume-title":"Proc. ICLR","author":"Zhili"},{"key":"ref27","article-title":"Higher layers need more LoRA experts","author":"Gao","year":"2024","journal-title":"arXiv:2402.08562"},{"key":"ref28","article-title":"SiRA: Sparse mixture of low rank adaptation","author":"Zhu","year":"2023","journal-title":"arXiv:2311.09179"},{"key":"ref29","article-title":"MultiLoRA: Democratizing LoRA for better multi-task learning","author":"Wang","year":"2023","journal-title":"arXiv:2311.11501"},{"key":"ref30","article-title":"The art of balancing: Revolutionizing mixture of experts for maintaining world knowledge in language model alignment","author":"Dou","year":"2023","journal-title":"arXiv:2312.09979"},{"key":"ref31","article-title":"MoRAL: MoE augmented LoRA for LLMs\u2019 lifelong learning","author":"Yang","year":"2024","journal-title":"arXiv:2402.11260"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19809-0_36"},{"key":"ref33","article-title":"Continual LLaVA: Continual instruction tuning in large vision-language models","author":"Cao","year":"2024","journal-title":"arXiv:2411.02564"},{"key":"ref34","article-title":"Ensembles of low-rank expert adapters","volume-title":"Proc. 13th Int. Conf. Learn. Representations","author":"Li"},{"key":"ref35","article-title":"Merging LoRAs like playing LEGO: Pushing the modularity of LoRA to extremes through rank-wise clustering","volume-title":"Proc. 13th Int. Conf. Learn. Represent.","author":"Zhao"},{"key":"ref36","article-title":"MoELoRA: An MOE-based parameter efficient fine-tuning method for multi-task medical applications","author":"Liu","year":"2023","journal-title":"arXiv:2310.18339"},{"key":"ref37","article-title":"MixLoRA: Enhancing large language models fine-tuning with LoRA based mixture of experts","author":"Li","year":"2024","journal-title":"arXiv:2404.15159"},{"key":"ref38","article-title":"LIMA: Less is more for alignment","author":"Zhou","year":"2023","journal-title":"arXiv:2305.11206"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1991.3.1.79"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.2.181"},{"key":"ref41","article-title":"Outrageously large neural networks: The sparsely-gated mixture-of-experts layer","author":"Shazeer","year":"2017","journal-title":"arxiv:1701.06538"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.emnlp-main.243"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2024.3367329"},{"key":"ref44","article-title":"Adapt-${\\infty}$\n: Scalable continual multimodal instruction tuning via dynamic data selection","volume-title":"Proc. 13th Int. Conf. Learn. Represent.","author":"Maharana"},{"key":"ref45","article-title":"LLaVA-c: Continual improved visual instruction tuning","author":"Liu","year":"2025","journal-title":"arXiv:2506.08666"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.03762"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.1982.1056489"},{"key":"ref48","first-page":"8748","article-title":"Learning transferable visual models from natural language supervision","volume-title":"Proc. Int. Conf. Mach. Learn.","volume":"139","author":"Radford"},{"key":"ref49","article-title":"BLIP-2: Bootstrapping language-image pre-training with frozen image encoders and large language models","author":"Li","year":"2023","journal-title":"arxiv:2301.12597"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-58536-5_44"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.670"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00331"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-20074-8_9"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.2019.00156"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00166"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00686"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00566"},{"key":"ref59","article-title":"IconQA: A new benchmark for abstract diagram understanding and visual language reasoning","author":"Lu","year":"2021","journal-title":"arxiv:2110.13214"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.121"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00851"},{"key":"ref62","article-title":"Learn to explain: Multimodal reasoning via thought chains for science question answering","author":"Lu","year":"2022","journal-title":"arxiv:2209.09513"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1145\/3123266.3123427"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV48922.2021.00171"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7299087"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D19-1410"},{"key":"ref67","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2014","journal-title":"arXiv:1412.6980"},{"key":"ref68","first-page":"14702","article-title":"Exploring the benefits of training expert language models over instruction tuning","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Jang"},{"issue":"86","key":"ref69","first-page":"2579","article-title":"Visualizing data using t-SNE","volume":"9","author":"Maaten","year":"2008","journal-title":"JMLR"},{"key":"ref70","article-title":"Qwen2-VL: Enhancing vision-language model\u2019s perception of the world at any resolution","author":"Wang","year":"2024","journal-title":"arXiv:2409.12191"},{"key":"ref71","article-title":"G-LLaVA: Solving geometric problem with multi-modal large language model","author":"Gao","year":"2023","journal-title":"arXiv:2312.11370"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1038\/sdata.2018.251"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/ISBI48211.2021.9434010"},{"key":"ref74","article-title":"PathVQA: 30000 questions for medical visual question answering","author":"He","year":"2020","journal-title":"arXiv:2003.10286"},{"key":"ref75","article-title":"MME: A comprehensive evaluation benchmark for multimodal large language models","author":"Fu","year":"2023","journal-title":"arXiv:2306.13394"},{"key":"ref76","article-title":"MMBench: Is your multi-modal model an all-around player?","author":"Liu","year":"2023","journal-title":"arXiv:2307.06281"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2021.findings-acl.46"},{"key":"ref78","article-title":"LLaVA-Med: Training a large language-and-vision assistant for biomedicine in one day","author":"Li","year":"2023","journal-title":"arXiv:2306.00890"},{"key":"ref79","article-title":"SODA10M: Towards large-scale object detection benchmark for autonomous driving","author":"Han","year":"2021","journal-title":"arXiv:2106.11118"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-19839-7_24"},{"key":"ref81","doi-asserted-by":"publisher","DOI":"10.5555\/1953048.2078195"}],"container-title":["IEEE Transactions on Image Processing"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/83\/11355710\/11480879.pdf?arnumber=11480879","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,20]],"date-time":"2026-04-20T20:07:03Z","timestamp":1776715623000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11480879\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":81,"URL":"https:\/\/doi.org\/10.1109\/tip.2026.3678087","relation":{},"ISSN":["1057-7149","1941-0042"],"issn-type":[{"value":"1057-7149","type":"print"},{"value":"1941-0042","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}