Sign In

Communications of the ACM

ACM Opinion

Evolution Strategies as a Scalable Alternative to Reinforcement Learning

Evolution strategies

Credit: OpenAI

We've discovered that evolution strategies (ES), an optimization technique that's been known for decades, rivals the performance of standard reinforcement learning (RL) techniques on modern RL benchmarks (e.g. Atari/MuJoCo), while overcoming many of RL's inconveniences.

From OpenAI
View Full Article


No entries found