제목 [학술세미나] SEED 세미나 안내
작성일 2018-04-20 08:20:06
내용 SEED 세미나 안내

We cordially invite you to attend the SEED seminar by Su-Yun Huang (Academia Sinica, Taiwan) on Friday, 27 Apr 2018@3-4PM (UTC+8 note time difference).

Title:  Integrating multiple random sketches for sufficient dimension reduction in large-p-small-n problems
Speaker:   Su-Yun Huang , Academia Sinica
Time:   15:00-16:00(Taiwan), Fri, 27 April 2018
Virtual seminar room: webconf.vc.dfn.de/optimization<http://webconf.vc.dfn.de/optimization>
Room Passcode:   seed

Abstract: Sufficient dimension reduction (SDR) is continuing an active research field nowadays. When estimating the central subspace (CS), inverse regression based SDR methods involve solving a generalized eigenvalue problem, which can be problematic under the large-p-small-n situation. In recent years, there are emerging new techniques in numerical linear algebra, called randomized algorithms or random sketching, for high dimensional and large scale problems. To overcome the large-p-small-n problem in SDR, we combine the idea of statistical inference with random sketching to propose a new SDR method, named integrated random-partition SDR (iRP-SDR). Our method consists of the following steps. (1) Randomly partition the covariates into subsets to construct an envelope subspace with low dimension. (2) Obtain a sketch estimate of the CS by applying conventional SDR method in the constructed envelope subspace. (3) Repeat the above two steps for multiple times and integrate these multiple sketches to form a final estimate of the CS. The advantageous performance of iRP-SDR is demonstrated via simulation studies and an EEG data analysis. (joint with Hung Hung, National Taiwan University)

The SEED seminar series is jointly organized by researchers from National University of Singapore, Zuse Institute Berlin, The Institute of Statistical Mathematics, Academia Sinica, University College London, and Seoul National University. SEED stands for Statistics maschinElEarning Datascience. Motivated by the availability of big complex data and the fast development of new techniques in machine learning and data science, SEED aims to provide an online research platform for seminars focusing on important and timely interdisciplinary research topics on Statistics, Machine learning, Data Science, Mathematics, Operation Research, Computer Science, and Engineering. The online seminar series are co-hosted and organised by several research institutes in different countries. The mission is to exchange research ideas, educate young researchers, and promote international research and education collaborations.For more information, please visit the SEED website https://seed.stat.nus.edu.sg/index.php
파일 -