Distributed Reinforcement Learning for Scalable High-Performance Policy Optimization | Towards Data Science

Leveraging massive parallelism, asynchronous updates, and multi-machine training to match and exceed human-level performance