5月4日 Massimiliano Pontil教授学术报告

发布时间:2016-05-03

题 目：The Benefit of Multitask Representation Learning

报告人：Massimiliano Pontil, Professor, University College London

时 间： 5月4日（周三） 上午10：30-11：30

地 点：数据科学与bat365在线中国登录入口（东校区）A201

报告摘要：
We discuss a general method to learn data representations from multiple tasks. We provide a justification for this method in both settings of multitask learning and learning-to-learn. The method is illustrated in detail in the special case of linear feature learning. Conditions on the theoretical advantage offered by multitask representation learning over independent tasks learning are established. In particular, focusing on the important example of halfspace learning, we derive the regime in which multitask representation learning is beneficial over independent task learning, as a function of the sample size, the number of tasks and the intrinsic data dimensionality. Other potential applications of our results include multitask feature learning in reproducing kernel Hilbert spaces and multilayer, deep networks.

报告人简介：
http://www0.cs.ucl.ac.uk/staff/M.Pontil/