Улучшение политики с учетом причинности

AbstractThe recent success of reinforcement learning (RL) in solving complex tasks is often attributed to its capacity to explore and exploit an environment. Sample efficiency is usually not an issue …...