Connecting Generative Adversarial Networks and Actor-Critic Methods

Abstract

Both generative adversarial networks (GAN) in unsupervised learning and actor-critic methods in reinforcement learning (RL) have gained a reputation for being difficult to optimize. Practitioners in both fields have amassed a large number of strategies to mitigate these instabilities and improve training. Here we show that GANs can be viewed as actor-critic methods in an environment where the actor cannot affect the reward. We review the strategies for stabilizing training for each class of models, both those that generalize between the two and those that are particular to that model. We also review a number of extensions to GANs and RL algorithms with even more complicated information flow. We hope that by highlighting this formal connection we will encourage both GAN and RL communities to develop general, scalable, and stable algorithms for multilevel optimization with deep networks, and to draw inspiration across communities.

Added comments on inverse reinforcement learning

References in corpus (3)

• Energy-based Generative Adversarial Network
• A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models
• On distinguishability criteria for estimating generative models

Cited by in corpus (9)

• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• A Connection between Generative Adversarial Networks, Inverse Reinforcement Learning, and Energy-Based Models
• On the Global Convergence of Actor-Critic: A Case for Linear Quadratic Regulator with Ergodic Cost
• Efficient Regret Minimization in Non-Convex Games
• TF-Replicator: Distributed Machine Learning for Researchers
• Associative Adversarial Networks
• ACtuAL: Actor-Critic Under Adversarial Learning
• Stable Training of DNN for Speech Enhancement based on Perceptually-Motivated Black-Box Cost Function
• Generative Adversarial Imagination for Sample Efficient Deep Reinforcement Learning