Skip to yearly menu bar Skip to main content


Understanding the Impact of Entropy on Policy Optimization

Zafarali Ahmed · Nicolas Le Roux · Mohammad Norouzi · Dale Schuurmans

Pacific Ballroom #29

Keywords: [ Theory and Algorithms ] [ Optimization ] [ Non-convex Optimization ] [ Deep Reinforcement Learning ] [ Deep Learning Theory ]


Entropy regularization is commonly used to improve policy optimization in reinforcement learning. It is believed to help with exploration by encouraging the selection of more stochastic policies. In this work, we analyze this claim using new visualizations of the optimization landscape based on randomly perturbing the loss function. We first show that even with access to the exact gradient, policy optimization is difficult due to the geometry of the objective function. We then qualitatively show that in some environments, a policy with higher entropy can make the optimization landscape smoother, thereby connecting local optima and enabling the use of larger learning rates. This paper presents new tools for understanding the optimization landscape, shows that policy entropy serves as a regularizer, and highlights the challenge of designing general-purpose policy optimization algorithms.

Live content is unavailable. Log in and register to view live content