You can call me Zebu (/zəˈbuː/, pronounced “zuh-BOO”).

I am a Ph.D. candidate in the Department of Computer Science and Engineering (CSE) at The Chinese University of Hong Kong (CUHK), under the supervision of Prof. John C.S. Lui (ACM/IEEE Fellow). Before this, I earned a B.Eng. in 2024 from the Advanced Class in Mathematics and Physics for Information Science at Huazhong University of Science and Technology (1/30).

I am currently a Visiting Scholar at the SOLAR Lab, University of Massachusetts Amherst, hosted by Prof. Mohammad Hajiesmaili.

Research Interest

TL;DR: Online Learning, Large Language Models, Video Streaming, Edge Computing.

My previous research focused on adjustable spatial encoding strategies for video analysis, particularly in adapting video quality to fluctuating bandwidth conditions using online feedback. This experience with the highly dynamic and unpredictable nature of video streaming directly sparked my interest in the core principles of continuous adaptation.

Consequently, my research focus has shifted to Online Learning for Large Language Models (LLMs), with the goal of enabling them to adapt beyond static training. I am interested in Diffusion LLMs for iterative response refinement toward more coherent answers. More broadly, I study how LLMs connect knowledge and reason by analogy, and how these abilities can help agents learn quickly from interaction.

Publications

(*equal contribution)

• [TMC] Online Outcome-Driven Configuration Selection with Adaptive Grouping for Live Video Streams
  Xiangxiang Dai*, Zeyu Zhang*, Ziyi Han, Xutong Liu, Peng Yang, John C.S. Lui.
  Accepted by IEEE Transactions on Mobile Computing (TMC), 2026.
  

• [ACM MM] AxiomVision: Accuracy-Guaranteed Adaptive Visual Model Selection for Perspective-Aware Video Analytics
  Xiangxiang Dai, Zeyu Zhang, Peng Yang, Yuedong Xu, Xutong Liu, John C.S. Lui.
  Accepted in the ACM Multimedia (MM), 2024. (1149/4385 = 26.2%, CCF A)
  [Link], [arXiv], [Poster], [Code], [ACM showcase on Kudos]
  

• [ICC] Bandwidth-Efficient Edge Video Analytics via Frame Partitioning and Quantization Optimization
  Chuqin Zhou, Peng Yang, Zeyu Zhang, Chengzhi Wang, Ning Zhang.
  Accepted in the IEEE International Conference on Communications (ICC), 2023.

Preprints

• [arXiv] Steering Frozen LLMs: Adaptive Social Alignment via Online Prompt Routing
  Zeyu Zhang, Xiangxiang Dai, Ziyi Han, Xutong Liu, John C.S. Lui

• [arXiv] HiLoRA: Adaptive Hierarchical LoRA Routing for Training-Free Domain Generalization
  Ziyi Han, Huanyu Wang, Zeyu Zhang, Xiangxiang Dai, Xutong Liu, John C.S. Lui

• [arXiv] Offline Clustering of Linear Bandits: Unlocking the Power of Clusters in Data-Limited Environments
  Jingyuan Liu*, Zeyu Zhang*, Xuchuang Wang, Xutong Liu, John C.S. Lui, Mohammad Hajiesmaili, Carlee Joe-Wong.
  

Internships and Research Experiences

• 2026.05-Present,   Visiting Scholar at the SOLAR Lab, University of Massachusetts Amherst, USA. Advisor: Prof. Mohammad Hajiesmaili.
• 2025.09-2026.04,   Research Intern at Hong Kong AI & Data Laboratory Limited (HKAI Lab), Hong Kong, China. Leader: Wu Huiwei.
• 2025.06-2025.08,   Research Intern at Huawei 2012 Lab, Hangzhou, China; Leader: Prof. Haibo Chen (ACM/IEEE Fellow).
• 2021.09-2023.06,   Research Intern of Prof. Peng Yang at Huazhong University of Science and Technology, Wuhan, China.

Selected Awards

• Huawei PhD Fellowship (2024-2028)

• National Scholarship (Oct. 2021, Oct. 2023)

• Merit Student (Oct. 2021, Oct. 2023)

• Outstanding Undergraduates in Term of Academic Performance (Oct. 2021)

• Top 1% in Alibaba Global Mathematics Competition (Apr. 2022).

Activities

Reviewer

• International Conference on Learning Representations (ICLR) 2025, 2026
• IEEE Internet of Things Journal (IoT-J)

Teaching Assistantships

• CSCI2040: Introduction to Python, CUHK, Fall 2024, 2025
• ENGG1110: Problem Solving By Programming, CUHK, Spring 2025, 2026