Deep Reinforcement Learning

Asynchronous Advantage Actor-Critic (A3C)

Distributed architecture where multiple agents train in parallel on copies of the environment, sampling uncorrelated trajectories and accelerating convergence.

← Geri