Kaishen Yuan

Kaishen Yuan is a PhD student in the AI Thrust at The Hong Kong University of Science and Technology (Guangzhou), having begun his doctoral studies in 2025 after completing an MS at Tianjin University. His research focuses on deep learning, facial action unit detection, and multimodal large language models, with work spanning emotional image generation, medical reasoning, and vision-language-action models. He is best known for AUFormer, a parameter-efficient vision transformer for facial action unit detection published at ECCV 2024, and has contributed to the DeltaVLA project.