Asynchronous deep reinforcement learning

摘要

Methods, systems and apparatuses include a computer program encoded on a computer storage medium for asynchronous deep reinforcement learning. One of the systems includes one or more computers configured to implement a plurality of workers, each worker being configured to operate independently of each other worker, and each worker is configured to operate in the deep neural network during training of the environment. Each Replica is associated with each Actor interacting with it. Aspects of the present specification have a technical effect of faster training of a neural network and/or a reduction in memory requirements for the training.