Collaborative Evolutionary Reinforcement Learning

Intel Researchers created a new approach to RL via Collaborative Evolutionary Reinforcement Learning (CERL) that combines policy gradient and evolution methods to optimize, exploit, and explore challenges.