Fanhu Zeng

I'm currently a Master student at Institute of Automation, Chinese Academy of Sciences (CASIA), supervised by Prof. Xu-Yao Zhang and Prof. Cheng-Lin Liu. My major is Pattern Recognition and Intelligent System.

Before that, I obtained my Bachlor degree from Nanjing University of Aeronautics and Astronautics (NUAA) majoring in Automation.

My research interests lie in:
1. Trustworthy AI

Uncertainty Estimation
Out-of-distribution Detection
Continual Learning

2. Multimodal Large Language models

Prompt Learning/LoRA
Downstream Application

3. Efficient AI

Model Merging
Token Compression/Model Acceleration
Image Compression

Particularly the up-to-date combination of Multimodal Large Language Models with Trustworthy AI like Hallucination, Continual Instruction Tuning and so on.

Email / Google Scholar / Github / LinkedIn

📣 I am looking for PhD opportunities for these directions starting in 25Fall, 26Spring or 26Fall. I would greatly appreciate it if you have any available positions or suggestions.

Education

M.S. @ Institute of Automation, Chinese Academy of Sciences, 2022-2025
Advisor: Prof. Xu-Yao Zhang and Prof. Cheng-Lin Liu
B.S. @ Nanjing University of Aeronautics and Astronautics, 2018-2022
Advisor: Prof. Xin Yang and Prof. Fuyang Chen

News

[2025.05] One paper (FCIT) is accepted to ICCV 2025. Congratulations to all my collaborators!
[2025.05] Two papers (HiDe-LLaVA, ChartEdit) are accepted to ACL 2025. Congratulations to all my collaborators!
[2025.01] One paper (Local-Prompt) is accepted to ICLR 2025.

Publications

* indicates equal contribution

	Federated Continual Instruction Tuning Haiyang Guo, Fanhu Zeng, Fei Zhu, Wenzhuo Liu, Da-Han Wang, Jian Xu, Xu-Yao Zhang, Cheng-Lin Liu International Conference on Computer Vision (ICCV), 2025 arXiv / Code
	HiDe-LLaVA: Hierarchical Decoupling for Continual Instruction Tuning of Multimodal Large Language Model Haiyang Guo, Fanhu Zeng, Ziwei Xiang, Fei Zhu, Da-Han Wang, Xu-Yao Zhang, Cheng-Lin Liu Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL), 2025 arXiv / Code
	Local-Prompt: Extensible Local Prompts for Few-Shot Out-of-Distribution Detection Fanhu Zeng, Zhen Cheng, Fei Zhu, Hongxin Wei, Xu-Yao Zhang The Thirteenth International Conference on Learning Representations (ICLR), 2025 Paper / arXiv / Code
	EventVAD: Training-Free Event-Aware Video Anomaly Detection Yihua Shao, Haojin He, Sijie Li, Siyu Chen, Xinwei Long, Fanhu Zeng, Yuxuan Fan, Muyang Zhang, Ziyang Yan, Ao Ma, Xiaochen Wang, Hao Tang, Yan Wang, Shuyan Li The 33rd ACM International Conference on Multimedia (ACM MM), 2025 arXiv / Code
	ChartEdit: How Far Are MLLMs From Automating Chart Analysis? Evaluating MLLMs’ Capability via Chart Editing Xuanle Zhao, Xuexin Liu, Haoyue Yang, Xianzhen Luo, Fanhu Zeng, Jianling Li, Qi Shi, Chi Chen Findings of the Association for Computational Linguistics (ACL)*, 2025 arXiv / Code
	M2M-TAG: Training-Free Many-to-Many Token Aggregation for Vision Transformer Acceleration Fanhu Zeng, Deli Yu Workshop on Machine Learning and Compression, Neural Information Processing Systems (NeurIPS), 2024 Paper

Preprints

	A Comprehensive Survey on Continual Learning in Generative Models Haiyang Guo, Fanhu Zeng, Fei Zhu, Jiayi Wang, Xukai Wang, Jingang Zhou, Hongbo Zhao, Wenzhuo Liu, Shijie Ma, Da-Han Wang, Xu-Yao Zhang, Cheng-Lin Liu arXiv / Code
	Token Reduction Should Go Beyond Efficiency in Generative Models – From Vision, Language to Multimodality Zhenglun Kong, Yize Li, Fanhu Zeng, Lei Xin, Shvat Messica, Xue Lin, Pu Zhao, Manolis Kellis, Hao Tang, Marinka Zitnik arXiv / Code
	TR-DQ: Time-Rotation Diffusion Quantization Yihua Shao, Deyang Lin, Fanhu Zeng, Minxi Yan, Muyang Zhang, Siyu Chen, Yuxuan Fan, Ziyang Yan, Haozhe Wang, Jingcai Guo, Yan Wang, Haotong Qin, Hao Tang arXiv
	Towards Efficient and General-Purpose Few-Shot Misclassification Detection for Vision-Language Models Fanhu Zeng, Zhen Cheng, Fei Zhu, Xu-Yao Zhang arXiv
	Parameter Efficient Merging for Multimodal Large Language Models with Complementary Parameter Adaptation Fanhu Zeng, Haiyang Guo, Fei Zhu, Li Shen, Hao Tang arXiv
	Dynamic Knowledge Consolidation for Rehearsal-Free Continual Learning Haiyang Guo, Fei Zhu, Fanhu Zeng, Bing Liu, Xu-Yao Zhang arXiv
	ModalPrompt: Dual-Modality Guided Prompt for Continual Learning of Large Multimodal Models Fanhu Zeng, Fei Zhu, Haiyang Guo, Xu-Yao Zhang, Cheng-Lin Liu arXiv
	PPT: Token Pruning and Pooling for Efficient Vision Transformers Xinjian Wu, Fanhu Zeng, Xiudong Wang, Xinghao Chen arXiv / Code

Academic Services

Conference Reviewer: EMNLP, NeurIPS, ICLR, CVPR, ICCV

Website Template