phbs
2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | 2011 |
Relaxed Wasserstein, with Applications to GANs and Beyond
2019-01-10 23:17:41
by Xin Guo, University of California

Wednesday, Dec 19, 2018 | 4:00pm-5:30pm | Room 333, HSBC Business School Building


Abstract


In this talk, I will review some well known distance measures that are widely used in statistics, machine learning and stochastic games. These include Bregman divergence, Wasserstein distance, and a recently proposed new divergence function names relaxed Wasserstein. I will review some of their important properties and implications in machine learning and optimization. We will then discuss the application to GANs, a central topic in machine learning, and its potential connection with mathematical finance.
COPYRIGHT @ 2013-2018 PHBS · SUPPORT BY ITS.PKUSZ.DEU.CN Guangdong-ICP prepared No.12081285