Continuous Image Recognition under Micromemory

Professor Weishi Zheng
Sun Yat-sen University, China

For a long time, we hope that deep learning models can continue to learn for new problems, new categories, new data, etc. However, due to the catastrophic forgetting problem, when the deep learning model is optimized for the new task, the classification performance of the original task will be seriously degraded. For this reason, many novel continuous learning algorithms have been proposed in recent years. On continuous learning, we have recently investigated this problem, mainly about how to use unlabeled data to solve the continuous learning modeling problem under small memory and how to use prompt modeling to solve the continuous learning problem under zero memory environment to quickly adapt to downstream tasks. We will present these recent explorations and look forward to discussing them with you.

Biosketch

Weishi Zheng is now a full Professor with Sun Yat-sen University. His research interests include person/object association and activity understanding, and the related weakly supervised/unsupervised and continuous learning machine learning algorithms. He has now published more than 200 papers, including more than 150 publications in main journals (TPAMI, IJCV, SIGGRAPH, TIP) and top conferences (ICCV, CVPR, ECCV, NeurIPS). He has ever served as area chairs of ICCV, CVPR, ECCV, BMVC, NeurIPS and etc. He is associate editors/on the editorial board of IEEE-TPAMI, Artificial Intelligence Journal, Pattern Recognition. He has ever joined Microsoft Research Asia Young Faculty Visiting Programme. He is a Cheung Kong Scholar Distinguished Professor, a recipient of the Excellent Young Scientists Fund of the National Natural Science Foundation of China, and a recipient of the Royal Society-Newton Advanced Fellowship of the United Kingdom.