Convergence Problems with Generative Adversarial Networks (GANs)

Abstract

Generative adversarial networks (GANs) are a novel approach to generative modelling, a task whose goal it is to learn a distribution of real data points. They have often proved difficult to train: GANs are unlike many techniques in machine learning, in that they are best described as a two-player game between a discriminator and generator. This has yielded both unreliability in the training process, and a general lack of understanding as to how GANs converge, and if so, to what. The purpose of this dissertation is to provide an account of the theory of GANs suitable for the mathematician, highlighting both positive and negative results. This involves identifying the problems when training GANs, and how topological and game-theoretic perspectives of GANs have contributed to our understanding and improved our techniques in recent years.

47 pages, 4 figures

References in corpus (7)

• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Towards Principled Methods for Training Generative Adversarial Networks
• Many Paths to Equilibrium: GANs Do Not Need to Decrease a Divergence At Every Step
• From optimal transport to generative modeling: the VEGAN cookbook
• On the Discrimination-Generalization Tradeoff in GANs
• Efficient Regret Minimization in Non-Convex Games
• A Geometric View of Optimal Transportation and Generative Model

Cited by in corpus (11)

• A survey on GANs for computer vision: Recent research, analysis and taxonomy
• Generalizing semi-supervised generative adversarial networks to regression using feature contrasting
• Action Robust Reinforcement Learning and Applications in Continuous Control
• Stabilizing Generative Adversarial Networks: A Survey
• The Inductive Bias of Restricted f-GANs
• Is Discriminator a Good Feature Extractor?
• HAD-Net: A Hierarchical Adversarial Knowledge Distillation Network for Improved Enhanced Tumour Segmentation Without Post-Contrast Images
• DAMIA: Leveraging Domain Adaptation as a Defense against Membership Inference Attacks
• Conditional Generative Modeling via Learning the Latent Space
• Generative Adversarial Network: Some Analytical Perspectives
• Convergence of GANs Training: A Game and Stochastic Control Methodology