Our finding continues the modern trend of achieving strong results with decades-old ideas. For example, in 2012, the “AlexNet” paper⁠(opens in a new window) showed how to design, scale and train convolutional neural networks (CNNs) to achieve extremely strong results on image recognition tasks, at a time when most researchers thought that CNNs were not a promising approach to computer vision. Similarly, in 2013, the Deep Q-Learning paper⁠(opens in a new window) showed how to combine Q-Learning with CNNs to successfully solve Atari games, reinvigorating RL as a research field with exciting experimental (rather than theoretical) results. Likewise, our work demonstrates that ES achieves strong performance on RL benchmarks, dispelling the common belief that ES methods are impossible to apply to high dimensional problems.

ES is easy to implement and scale. Running on a computing cluster of 80 machines and 1,440 CPU cores, our implementation is able to train a 3D MuJoCo humanoid walker in only 10 minutes (A3C on 32 cores takes about 10 hours). Using 720 cores we can also obtain comparable performance to A3C on Atari while cutting down the training time from 1 day to 1 hour.

In what follows, we’ll first briefly describe the conventional RL approach, contrast that with our ES approach, discuss the tradeoffs between ES and RL, and finally highlight some of our experiments.

Let’s briefly look at how RL works. Suppose we are given some environment (e.g. a game) that we’d like to train an agent on. To describe the behavior of the agent, we define a policy function (the brain of the agent), which computes how the agent should act in any given situation. In practice, the policy is usually a neural network that takes the current state of the game as an input and calculates the probability of taking any of the allowed actions. A typical policy function might have about 1,000,000 parameters, so our task comes down to finding the precise setting of these parameters such that the policy plays well (i.e. wins a lot of games).