Finding Mixed Nash Equilibria of Generative Adversarial NetworksDownload PDF

27 Sept 2018 (modified: 22 Oct 2023)ICLR 2019 Conference Blind SubmissionReaders: Everyone
Abstract: We reconsider the training objective of Generative Adversarial Networks (GANs) from the mixed Nash Equilibria (NE) perspective. Inspired by the classical prox methods, we develop a novel algorithmic framework for GANs via an infinite-dimensional two-player game and prove rigorous convergence rates to the mixed NE. We then propose a principled procedure to reduce our novel prox methods to simple sampling routines, leading to practically efficient algorithms. Finally, we provide experimental evidence that our approach outperforms methods that seek pure strategy equilibria, such as SGD, Adam, and RMSProp, both in speed and quality.
Keywords: GANs, mixed Nash equilibrium, mirror descent, sampling
Data: [LSUN](https://paperswithcode.com/dataset/lsun)
Community Implementations: [![CatalyzeX](/images/catalyzex_icon.svg) 1 code implementation](https://www.catalyzex.com/paper/arxiv:1811.02002/code)
14 Replies

Loading