[행사/세미나] 인공지능대학원 전문가 초청 세미나(POSTECH 곽수하 교수, 10/4 14:00~)
- 인공지능학과
- 조회수24560
- 2019-09-23
일시: 10월 4일 금요일 14:00 ~ 15:30
장소: 반도체관 400126호
Title: Weakly Supervised Learning for Semi-automatic Labeling of Visual Data
Speaker: Suha Kwak (곽수하 교수, POSTECH)
Homepage: http://cvlab.postech.ac.kr/~suhakwak/
Biography. Suha Kwak is an assistant professor in Computer Science and Engineering at POSTECH, and a faculty member of POSTECH Computer Vision Lab. Before that, he spent one and half years on the faculty of the Department of Information and Communication Engineering at DGIST. He did a post-doc with Ivan Laptev and Jean Ponce in the WILLOW team at the Department of Computer Science of the École Normale Supérieure and Inria Paris. He completed my BS and PhD in 2007 and 2014, respectively, both at POSTECH, where he was advised by Prof. Joon Hee Han and Prof. Bohyung Han. He is primarily interested in techniques for learning visual recognition models with less or no human supervision, such as weakly supervised learning, semi-supervised learning, webly supervised learning, transfer learning, and domain adaptation. His current research interests also include metric learning and its applications, visual recognition in extreme conditions, and medical image analysis.
Abstract. Supervised learning of Convolutional Neural Networks (CNNs) has driven recent advances in visual recognition. Due to the data-hungry nature of deep CNNs, this approach demands an enormous number of training images with groundtruth labels, which are given by hand in general. However, manual annotation of the labels is prohibitively time-consuming for high-level visual recognition tasks like semantic segmentation, which results in existing datasets limited in terms of both class diversity and the amount of labeled data. It is thus not straightforward to learn high-level visual recognition models that can handle diverse object classes in the real world. As a way to alleviate this issue, this talk suggests weakly supervised learning that adopts weaker yet less expensive and readily available labels as supervision. As well as the definition, motivations, and challenges of weakly supervised learning, recent research of our group on the topic will be introduced in this talk.