Speech Title: Recent Progress on Weakly-Supervised Learning for Imperfect Data
Traditional supervised machine learning algorithms usually require sufficient, accurate, and definite training data during training. However, in many practical situations, due to a variety of objective limitations, the perfectness of training data is difficult to guarantee. Therefore, in this talk, I will mainly introduce the latest progress of our group in the field of weakly supervised learning. Focusing on common practical problems such as insufficient supervision, inaccurate supervision, and indefinite supervision, I will mainly introduce the corresponding machine learning algorithms including semi-supervised learning, PU learning, label noise learning, partial label learning, etc. The related theoretical results and empirical results will also be introduced.
Chen Gong received his B.E. degree from East China University of Science and Technology (ECUST) in 2010, and a dual doctoral degree from Shanghai Jiao Tong University (SJTU) and University of Technology Sydney (UTS) in 2016 and 2017, respectively. Currently, he is a full professor in the School of Computer Science and Engineering, Nanjing University of Science and Technology. He has published more than 90 technical papers at prominent journals and conferences such as IEEE T-PAMI, IEEE T-NNLS, IEEE T-IP, IEEE T-CYB, ICML, NeurIPS, CVPR, AAAI, IJCAI, ICDM, etc, and also holds six granted invention patents. He also serves as the reviewer for more than 20 international journals such as AIJ, JMLR, IEEE T-PAMI, IEEE T-NNLS, IEEE T-IP, IEEE T-KDE, and also the SPC/PC member of several top-tier conferences such as ICML, NeurIPS, CVPR, AAAI, IJCAI, ICDM, etc. He received the "Wu Wen-Jun AI Excellent Youth Scholar Award", "Young Elite Scientists Sponsorship Program" of China Association for Science and Technology, "Hong Kong Scholar", and "Excellent Doctoral Dissertation award" by Shanghai Jiao Tong University (SJTU) and Chinese Association for Artificial Intelligence (CAAI).