Dahua Lin

Dahua Lin is an Associate Professor (by courtesy) in the Department of Information Engineering at The Chinese University of Hong Kong. His research interests include deep learning technologies in large-scale applications, high-level visual understanding in connection with linguistic analysis, efficient modeling and analytics of videos and movies, and the use of deep networks in general probabilistic inference. Lin has published over 120 papers on top conferences and journals, including CVPR, ICCV, ECCV, ICML, NeurIPS, and T-PAMI. He has also supervised or co-supervised CUHK teams in international competitions, winning multiple awards in ImageNet, ActivityNet, and COCO. Additionally, he has served on the editorial board of the International Journal of Computer Vision (IJCV) and as an area chair for several conferences.

Dahua Lin is a researcher at The Chinese University of Hong Kong, specializing in computer vision, machine learning, probabilistic inference, and Bayesian nonparametrics. His work focuses on advancing deep learning techniques for action recognition, object detection, and video generation. Lin has contributed to the development of frameworks such as MMDetection and MMSegmentation, which are widely used in computer vision tasks. His research includes temporal segment networks, spatial-temporal graph convolutional networks, and methods for improving multi-modal models through benchmarking and scaling. Lin's publications emphasize the integration of probabilistic models with deep learning to enhance performance in complex vision tasks.