The surge in interest in reinforcement learning is due to the fact that it revolutionizes automation by learning the optimal actions to take in an environment in order to maximize the notion of cumulative reward.
PyTorch 1.x Reinforcement Learning Cookbook introduces you to important reinforcement learning concepts and implementations of algorithms in PyTorch. Each chapter of the book walks you through a different type of reinforcement learning method and its industry-adopted applications. With the help of recipes that contain real-world examples, you will find it intriguing to enhance your knowledge and proficiency of reinforcement learning techniques in areas such as dynamic programming, Monte Carlo methods, temporal difference and Q-learning, multi-armed bandit, function approximation, deep Q-Networks, and policy gradients—they are no more obscure than you thought. Interesting and easy-to-follow examples, such as Atari games, Blackjack, Gridworld environments, internet advertising, Mountain Car, and Flappy Bird, will keep you interested until you reach your goal.
By the end of this book, you will have mastered the implementation of popular reinforcement learning algorithms and learned the best practices of applying reinforcement learning techniques to solve other real-world problems.