Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation

by   Yuhuai Wu, et al.

In this work, we propose to apply trust region optimization to deep reinforcement learning using a recently proposed Kronecker-factored approximation to the curvature. We extend the framework of natural policy gradient and propose to optimize both the actor and the critic using Kronecker-factored approximate curvature (K-FAC) with trust region; hence we call our method Actor Critic using Kronecker-Factored Trust Region (ACKTR). To the best of our knowledge, this is the first scalable trust region natural gradient method for actor-critic methods. It is also a method that learns non-trivial tasks in continuous control as well as discrete control policies directly from raw pixel inputs. We tested our approach across discrete domains in Atari games as well as continuous domains in the MuJoCo environment. With the proposed methods, we are able to achieve higher rewards and a 2- to 3-fold improvement in sample efficiency on average, compared to previous state-of-the-art on-policy actor-critic methods. Code is available at


page 1

page 2

page 3

page 4


Recursive Least Squares Advantage Actor-Critic Algorithms

As an important algorithm in deep reinforcement learning, advantage acto...

Curiosity-Driven Recommendation Strategy for Adaptive Learning via Deep Reinforcement Learning

The design of recommendations strategies in the adaptive learning system...

Exploiting the sign of the advantage function to learn deterministic policies in continuous domains

In the context of learning deterministic policies in continuous domains,...

GRAC: Self-Guided and Self-Regularized Actor-Critic

Deep reinforcement learning (DRL) algorithms have successfully been demo...

Using Soft Actor-Critic for Low-Level UAV Control

Unmanned Aerial Vehicles (UAVs), or drones, have recently been used in s...

Gradient Monitored Reinforcement Learning

This paper presents a novel neural network training approach for faster ...

Reinforcement Learning for Mixed-Integer Problems Based on MPC

Model Predictive Control has been recently proposed as policy approximat...