三、主 题：Generative Adversarial Networks as Variational Training of Energy Based Models
We study deep generative models for effective unsupervised learning. We propose VGAN, which works by minimizing a variational lower bound of the negative log likelihood (NLL) of an energy based model (EBM), where the model density p(x) is approximated by a variational distribution q(x) that is easy to sample from. The training of VGAN takes a two step procedure: given p(x), q(x) is updated to maximize the lower bound; p(x) is then updated one step with samples drawn from q(x) to decrease the lower bound. VGAN is inspired by the generative adversarial networks (GANs), where p(x) corresponds to the discriminator and q(x) corresponds to the generator, but with several notable differences. We hence name our model variational GANs (VGANs). VGAN provides a practical solution to training deep EBMs in high dimensional space, by eliminating the need of MCMC sampling. From this view, we are also able to identify causes to the difficulty of training GANs and propose viable solutions.
Shuangfei Zhai is currently a final year Ph.D student in Multimedia Research Lab, Department of Computer Science, Binghamton Univeristy, SUNY, where he work with Prof. Zhongfei (Mark) Zhang. Before coming to Binghamton University, he obtained my B.E. in Electronic Engineering and Information Science in University of Science and Technology of China, Hefei in 2010. He was a master student in Chinese Academy of Sciences during 2010-2012.
He is now on job market looking for a Machine Learning/Deep Learning Research Scientist position.
编辑：林坤 / 审核：林坤 / 发布者：林坤