|제목||인공지능 대학원 전문가 초청 세미나(12/4 14:00~)|
아래와 같이 인공지능대학원에서 전문가 초청세미나를 주최하오니 관심있는 학생들의 많은 참여를 바랍니다.
- 아 래 -
● 시간: 12월 4일(수) 14:00 ~ 15:00
● 장소: 화학관 330110호
Title: Convolutional Neural Networks (CNN) and Successive Subspace Learning (SSL): Similarities and Differences
Speaker: Dr. C.-C. Jay Kuo, University of Southern California
Convolutional Neural Networks (CNN) and Successive Subspace Learning (SSL): Similarities and Differences
C.-C. Jay Kuo
University of Southern California
It is common to project signals from a high-dimensional space to a low-dimension space for ease of signal analysis and manipulation. This technique is commonly known as the subspace method. One famous example is the principal component analysis (PCA). Being motivated by interpretable convolutional neural networks (CNNs), we revisit the classical subspace method and propose a new methodology, called successive subspace learning (SSL), where we cascade multiple subspace learning stages so as to capture features from gradually expanding neighborhoods. The SSL system contains four main concepts: 1) successive local-to-global subspace expansion, 2) subspace approximation, 3) label-assisted dimension reduction, 4) aggregation and classification. The SSL is mathematically transparent. The parameters of an SSL system are determined in a feedforward manner. Since no backpropagation is needed, its training complexity is much lower than that for CNNs. Two concrete examples will be used to illustrate the SSL idea: 1) the PixelHop system for 2D image object classification and 2) the PointHop system for 3D point cloud classification. The SSL solutions achieve comparable performance of state-of-the-art CNN solutions yet at much lower complexity. Future research topics in SSL will be pointed out.
Dr. C.-C. Jay Kuo received his Ph.D. degree from the Massachusetts Institute of Technology in 1987. He is now with the University of Southern California (USC) as Director of the Media Communications Laboratory and Distinguished Professor of Electrical Engineering and Computer Science. His research interests are in the areas of media processing, compression and understanding. Dr. Kuo was the Editor-in-Chief for the IEEE Trans. on Information Forensics and Security in 2012-2014. Dr. Kuo is a Fellow of AAAS, IEEE and SPIE. He has guided 150 students to their Ph.D. degrees and supervised 30 postdoctoral research fellows. Dr. Kuo is a co-author of 280 journal papers, 920 conference papers and 14 books. Dr. Kuo received the 2016 IEEE Computer Society Taylor L. Booth Education Award, the 2016 IEEE Circuits and Systems Society John Choma Education Award, the 2016 IS&T Raymond C. Bowman Award, the 2017 IEEE Leon K. Kirchmayer Graduate Teaching Award, the 2017 IEEE Signal Processing Society Education Award, and the 2019 IEEE Computer Society Edward J. McCluskey Technical Achievement Award.