C-GAIL: Stabilizing Generative Adversarial Imitation Learning With Control Theory

Abstract

Generative Adversarial Imitation Learning (GAIL) trains a generative policy to mimic a demonstrator. It uses on-policy Reinforcement Learning (RL) to optimize a reward signal derived from a GAN-like discriminator. A major drawback of GAIL is its training instability - it inherits the complex training dynamics of GANs, and the distribution shift introduced by RL. This can cause oscillations during training, harming its sample efficiency and final policy performance. Recent work has shown that control theory can help with the convergence of a GAN's training. This paper extends this line of work, conducting a control-theoretic analysis of GAIL and deriving a novel controller that not only pushes GAIL to the desired equilibrium but also achieves asymptotic stability in a 'one-step' setting. Based on this, we propose a practical algorithm 'Controlled-GAIL' (C-GAIL). On MuJoCo tasks, our controlled variant is able to speed up the rate of convergence, reduce the range of oscillation and match

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyluo2024c

Related papers

• When Will Generative Adversarial Imitation Learning Algorithms Attain Global Convergence (2020)0.00
• Non-adversarial Imitation Learning And Its Connections To Adversarial Methods (2020)0.00
• \(f\)-gail: Learning \(f\)-divergence For Generative Adversarial Imitation Learning (2020)0.00
• A Pragmatic Look At Deep Imitation Learning (2021)0.00
• Provably Efficient Generative Adversarial Imitation Learning For Online And Offline Setting With Linear Function Approximation (2021)0.00
• Lipschitzness Is All You Need To Tame Off-policy Generative Adversarial Imitation Learning (2020)7.81
• Augmenting GAIL With BC For Sample Efficient Imitation Learning (2020)0.00
• GIFT: Global Stabilisation Via Intrinsic Fine Tuning (2026)0.00