Агенты ИИ | AGI_and_RL

A New Framework for Multi-Agent Reinforcement Learning –

Centralized Training and Exploration with Decentralized

Execution via Policy Distillation https://arxiv.org/pdf/1910.09152.pdf