1 paper
Yeu-Chern Harn, Zhenghao Chen, Vladimir Jojic
In this work, we propose a composition/decomposition framework for adversarially training generative models on composed data - data where each sample can be thought of as being con…