- 제목
- 세미나 [09/26] A randomly Weighted Neural Network for Mutual Information Estimation with application to Time-series Change-points Detection
- 작성일
- 2016.09.22
- 작성자
- 최고관리자
- 게시글 내용
-
< BK21 플러스 BEST 정보기술 사업단 세미나 개최 안내 >
개최일시 : 2016년 09월 26일 월요일 10:00 ~ 11:00
개최장소 : 제 3공학관 379C호
세미나 제목 : A randomly Weighted Neural Network for Mutual Information Estimation with application to Time-series Change-points Detection
발표초록 :
In this talk, we propose an efficient parameter tuning-free Mutual Information (MI) estimator in a form of a Radial Basis Function (RBF) network. The input layer of the proposed network propagates a sample pair of two random variables to the hidden layer. The propagated samples are then transformed by a set of Gaussian RBF kernels with randomly determined kernel centers and widths similar to that in an extreme learning machine. The output layer adopts a linear weighting scheme which can be analytically estimated. Our empirical results show that the proposed estimator outperforms the competing state-of-the-art MI estimators in terms of computational efficiency while showing the comparable estimation accuracy performance. Moreover, the proposed model achieves promising results in an application study of time-series change-points detection and driving stress analysis.
강연자 : 오범석, Research Fellow / Nanyang Technological University, Singapore
초청자 : 전기전자공학과 교수 토카안