Haokun Lin

Haokun Lin

Ph.D. Student

Hi, I'm Haokun Lin (ๆž—ๆตฉๅค) ๐Ÿป

Iโ€™m a Ph.D. candidate at New Laboratory of Pattern Recognition (NLPR), Institute of Automation, Chinese Academy of Sciences under the supervision of Prof. Zhenan Sun. Iโ€™m also a joint Ph.D. candidate at Department of Computer Science, City University of Hong Kong, working with Prof. Ying Wei and Prof. Zhichao Lu. Before joining CASIA, I received my B.Eng. in Software Engineering from Huazhong University of Science and Technology in 2021.

My research interests include Multi-modal Learning, Large Language/Vision Models, and Efficient Deep Learning.

๐Ÿ‘‹๐Ÿ‘‹๐Ÿ‘‹ If youโ€™re interested in my work, please feel free to reach out for discussions or collaborations!

Contact me via:
๐Ÿ“ง Mail: haokun.lin[AT]cripac.ia.ac.cn or haokunlin2-c[AT]my.cityu.edu.hk

๐ŸŒˆ What's new:

  • [05/2026] ๐Ÿš€ Award: Honored to be selected as Silver Reviewer for ICML'26!
  • [05/2026] ๐ŸŽ‰ ICML'26: "MedREK: Retrieval-Based Editing for Medical LLMs with Key-Aware Prompts." [Code/PDF]
  • [05/2026] ๐ŸŽ‰ ICML'26: "Concept-Guided Tokenization: Closing the Gap Between Reconstruction and Generation."
  • [04/2026] ๐ŸŽ‰ Four papers are accepted to ACL'26, IJCNN'26, ICMR'26 and Neurocomputing.
  • [02/2026] ๐ŸŽ‰ CVPR'26: "QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Models." [Code/PDF]
  • [01/2026] ๐Ÿ“œ Preprint: "Efficient Diffusion Language Models: A Comprehensive Survey." [Repo/PDF]
  • [11/2025] ๐ŸŽ‰ MIR: "Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs." [Code/PDF]
  • [11/2025] ๐ŸŽ‰ ResponsibleFM @ NeurIPS 2025: "MedREK: Retrieval-Based Editing for Medical LLMs with Key-Aware Prompts." [Code/PDF]
  • [11/2025] ๐Ÿš€ Award: Delighted to have received the National Scholarship at UCAS! Grateful to my supervisors!
  • [08/2025] ๐Ÿ“œ Preprint: "LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Text-to-Image Generation." [PDF]
  • [06/2025] ๐ŸŽ‰ ICCV'25: "DOGR: Towards Versatile Visual Document Grounding and Referring." [Code/PDF]
  • [05/2025] ๐Ÿ“œ Preprint: "TokLIP: Marry Visual Tokens to CLIP for Multimodal Comprehension and Generation." [Code/PDF]
  • [02/2025] ๐ŸŽ‰ TMM: "Scale Up Composed Image Retrieval Learning via Modification Text Generatio." [PDF]
  • [01/2025] ๐ŸŽ‰ ICLR'25: "Image-level Memorization Detection via Inversion-based Inference Perturbation." [PDF]
  • [11/2024] ๐Ÿš€ Award: Delighted to have received the First Prize in the 2024 Graduate Academic Forum at UCAS!
  • [11/2024] ๐Ÿš€ Award: Honored to be selected as a Top Reviewer at NeurIPS 2024!
  • [09/2024] ๐ŸŽ‰ NeurIPS'24 Oral: "DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs." Big Congs! ๐Ÿ”ฅ๐Ÿ”ฅ๐Ÿ”ฅ [Code/PDF]
  • [07/2024] ๐ŸŽ‰ ECCV'24: "MATHVERSE: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?" [Code/PDF]
  • [05/2024] ๐ŸŽ‰ ACL'24 Findings: "IntactKV: Improving Large Language Model Quantization by Keeping Pivot Tokens Intact." [Code/PDF]
  • [02/2024] ๐ŸŽ‰ CVPR'24: "MoPE-CLIP: Structured Pruning for Efficient Vision-Language Models with Module-wise Pruning Error Metric." [PDF]
  • [01/2024] ๐ŸŽ‰ ICLR'24: "Plug-and-Play: An Efficient Post-training Pruning Method for Large Language Models." [Code/PDF]
  • [03/2022] ๐ŸŽ“ Starting Joint Ph.D.@CityU: I will join Prof. Ying Wei's group at CityU in 2022 Fall!
  • [09/2021] ๐ŸŽ“ Starting Ph.D.@CASIA: I will join Prof. Zhenan Sun's group at NLPR, CASIA in 2021 Fall!
  • [06/2021] ๐ŸŽ“ Graduation@HUST: Recieved my Bachelor's Degree from Huazhong University of Science and Technology with Honorary degree.

๐ŸŽ“ Selected Publications (Google Scholar)

(*: co-first author; ^: corresponding author; #: Project Lead)

DuQuant: Distributing Outliers via Dual Transformation Makes Stronger Quantized LLMs.
Haokun Lin*, Haobo Xu*, Yichen Wu*, Jingzhi Cui, Yingtao Zhang, Linzhan Mou, Linqi Song, Zhenan Sun^, Ying Wei^,
in 38th Conference on Neural Information Processing Systems (NeurIPS 2024 Oral).
[PDF] [arXiv] [Project] [Github] [QbitAI/้‡ๅญไฝ] [bibtex]
MoPE-CLIP: Structured Pruning for Efficient Vision-Language Models with Module-wise Pruning Error Metric.
Haokun Lin, Haoli Bai, Zhili Liu, Lu Hou, Muyi Sun, Linqi Song, Ying Wei^, Zhenan Sun^,
in IEEE / CVF Computer Vision and Pattern Recognition Conference 2024 (CVPR 2024).
[PDF] [arXiv] [bibtex]
TokLIP: Marry Visual Tokens to CLIP for Multimodal Comprehension and Generation.
Haokun Lin*, Teng Wang*, Yixiao Ge^, Yuying Ge, Zhichao Lu, Ying Wei, Qingfu Zhang, Zhenan Sun, Ying Shan,
Preprint.
[PDF] [arXiv] [Github] [HuggingFace] [QbitAI/้‡ๅญไฝ] [bibtex]
Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs.
Haokun Lin*, Haobo Xu*, Yichen Wu, Ziyu Guo, Renrui Zhang, Zhichao Lu, Ying Wei, Qingfu Zhang, Zhenan Sun,
in Machine Intelligence Research, 2025.
[PDF] [arXiv] [Github] [bibtex]
Efficient Diffusion Language Models: A Comprehensive Survey.
Haokun Lin*#, Xinle Jia*, Shaozhen Liu*, Shujun Xia*, Weitao Huang*, Haobo Xu, Junyang Li, Yicheng Xiao, Xingrun Xing, Ziyu Guo, Renrui Zhang, Qi Li, Yichen Wu, Renzhen Wang, Xiaojuan Qi, Caifeng Shan, Hongsheng Li, Zhenan Sun,
Preprint.
[PDF] [TechXriv] [Github] [Synced/ๆœบๅ™จไน‹ๅฟƒ] [bibtex]
MedREK: Retrieval-Based Editing for Medical LLMs with Key-Aware Prompts.
Shujun Xia*, Haokun Lin#*, Yichen Wu^, Yinan Zhou, Zixuan Li, Zhongwei Wan, Xingrun Xing, Yefeng Zheng, Xiang Li, Caifeng Shan, Zhenan Sun, Quanzheng Li^,
in Forty-Third International Conference on Machine Learning (ICML 2026).
[PDF] [arXiv] [Github] [bibtex]
Image-level Memorization Detection via Inversion-based Inference Perturbation.
Yue Jiang*, Haokun Lin*, Yang Bai, Bo Peng, Zhili Liu, Yueming Lyu, Yong Yang, Xing Zheng, Jing Dong,
in 13th International Conference on Learning Representations (ICLR 2025).
[PDF] [bibtex]
QuantVLA: Scale-Calibrated Post-Training Quantization for Vision-Language-Action Models.
Jingxuan Zhang*, Yunta Hsieh*, Zhongwei Wan, Haokun Lin, Xin Wang, Ziqi Wang, Yingtie Lei, Mi Zhang,
in IEEE / CVF Computer Vision and Pattern Recognition Conference 2026 (CVPR 2026).
[PDF] [arXiv] [Project] [Github] [bibtex]
DOGR: Towards Versatile Visual Document Grounding and Referring.
Yinan Zhou*, Yuxin Chen*, Haokun Lin, Yichen Wu, Shuyu Yang, Zhongang Qi, Chen Ma, Li Zhu, Ying Shan,
in IEEE / CVF International Conference on Computer Vision 2025 (ICCV 2025).
[PDF] [arXiv] [Github] [bibtex]
MATHVERSE: Does Your Multi-modal LLM Truly See the Diagrams in Visual Math Problems?
Renrui Zhang*, Dongzhi Jiang*, Yichi Zhang*, Haokun Lin, Ziyu Guo, Pengshuo Qiu, Aojun Zhou, Pan Lu, Kai-Wei Chang, Peng Gao, Hongsheng Li,
in 18th European Conference on Computer Vision (ECCV 2024).
[PDF] [arXiv] [Project] [Github] [Dataset] [Synced/ๆœบๅ™จไน‹ๅฟƒ] [bibtex]
Plug-and-Play: An Efficient Post-training Pruning Method for Large Language Models.
Yingtao Zhang, Haoli Bai, Haokun Lin, Jialin Zhao, Lu Hou, Carlo Vittorio Cannistraci,
in 12th International Conference on Learning Representations (ICLR 2024).
[PDF] [OpenReview] [Github] [bibtex]
IntactKV: Improving Large Language Model Quantization by Keeping Pivot Tokens Intact.
Ruikang Liu, Haoli Bai, Haokun Lin, Yuening Li, Han Gao, Zhengzhuo Xu, Lu Hou, Jun Yao, Chun Yuan,
in Findings of 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024 Findings)
[PDF] [arXiv] [Github] [bibtex]
LRQ-DiT: Log-Rotation Post-Training Quantization of Diffusion Transformers for Text-to-Image Generation.
Lianwei Yang*, Haokun Lin*, Tianchen Zhao*, Yichen Wu, Hongyu Zhu, Ruiqi Xie, Zhenan Sun, Yu Wang, Qingyi Gu,
Preprint.
[PDF] [arXiv] [Github] [bibtex]

๐Ÿ† Honors and Awards

๐ŸŽ– Services

๐Ÿ’ฌ Talks

Site Analytics