9月21日 中国科学院李启寨研究员学术报告

发布时间:2017-09-14浏览次数:140

报 告 人:李启寨 研究员(中国科学院)

报告题目:Bayesian Neural Networks for Selection of Anticancer Drug Response Genes

报告时间:2017年9月21号(星期四)上午 10:00

报告地点:静远楼1506报告厅

报告人简介:

    李启寨,中国科学院数学与系统科学研究院研究员,2001年于中国科学技术大学获学士学位,2006年于中国科学院研究生院获博士学位,2006-2009年在美国国家卫生健康研究院国家癌症研究所从事博士后研究,主要研究方向包括生物统计、统计遗传学、农业统计等。发表及接收发表论文85篇,目前担任Scientific Reports,PLoS One, Journal of Applied Statistics, Journal of Systems Science & Complexity等杂志的编委。获国家优秀青年科学基金、国际统计研究所推选会员、中国科学院卢嘉锡青年人才奖、中国工业与应用数学学会优秀青年学者奖、美国国家癌症研究所Fellow突出科研成果奖等。

报告摘要:

    Recent advances in high-throughput biotechnologies have provided an unprecedented opportunity for biomarker discovery, which can be cast as a variable selection problem. This problem is challenging due to three difficulties: an unknown functional form of the nonlinear system, variable selection consistency, and high-demanding computation. To circumvent them, we employ a feed-forward neural network to approximate the unknown nonlinear function, conduct structure selection for the neural network, which induces variable selection, by choosing appropriate prior distributions that lead to the consistency of variable selection, and implement the population stochastic approximation Monte Carlo algorithm on the OpenMP platform which provides a linear speedup for the simulation. The numerical results indicate that the proposed method can execute very fast on a multicore computer and work very well for identification of relevant variables for general high-dimensional nonlinear systems. The proposed method is successfully applied to selection of anticancer drug response genes for the drug sensitivity data collected in the cancer cell line encyclopedia study.