Xiao Xiang Zhu

Achieve state-of-the-art few-shot anomaly detection in medical imaging by dynamically weighting reference images based on their relevance to the query, enabling effective knowledge transfer from limited normal examples.

Achieve state-of-the-art image restoration across diverse image domains by adaptively composing task and domain prompts learned from multimodal LLMs.

Geospatial agents can now reason more effectively about satellite imagery thanks to a new framework that aligns models with verified multi-step tool interactions and explicit reasoning traces.