Yizhou Han

My principal research background lies in distributed and intelligent systems, where I focused on making machine learning workloads—such as those in computer vision and large language models—run efficiently and adaptively across heterogeneous cloud–edge–device environments.

In my previous research, I explored topics in federated continual learning, distributed optimization, and system-level coordination, applying these ideas to problems like resource scheduling, inference acceleration, and network optimization. One of my recent projects, Cluster-MoE, proposed an adaptive scheduling framework for asynchronous drift in federated learning, achieving notable efficiency improvements over existing approaches.

Recently, I have been exploring multimodal large language models (MLLMs) and their applications in Earth Observation (EO). I am particularly fascinated by how multimodal foundation models can integrate visual, spatial, and textual data for large-scale environmental and urban understanding. My goal is to build intelligent AI systems that bridge efficient system design and multimodal model capability, thereby enabling scalable and efficient AI applications.

Currently, I completed my Master study in Computer Science (with Distinction) at the University of St Andrews, where I was listed on the Dean’s List for academic excellence. Previously, I completed my B.S. in Computer Science at North China Electric Power University. I also have two years of experience in Urban Studies at the University of Glasgow, and three years of work experience in geoinformation and remote sensing, which helped me develop an interdisciplinary perspective on intelligent systems in earth and urban environments.