Ahmet Eray Kocaman - Exceeding Human Level Expertise Deep Reinforcement Learning

Описание к видео Ahmet Eray Kocaman - Exceeding Human Level Expertise Deep Reinforcement Learning

Abstract: Reinforcement learning (RL) is a machine learning approach where an artificial agent learns by interacting with its environment, similar to how a biological agent does. From accumulated experiences, the agent can optimize specific objectives, expressed as cumulative rewards. In this work fundamentals of reinforcement learning for a Markow Decision Process are discussed, including most notably the Q-learning. Since these traditional techniques become insufficient in high-dimensional environments, we introduced Deep Q-Learning, which integrates Q-learning with deep neural networks. Lastly, an application involving drone control, where performance surpasses that of human experts, is presented.

Комментарии

Информация по комментариям в разработке

Похожие видео