Reinforcement Learning with R

Rubén Oliva Ramos

  • 出版商: Packt Publishing
  • 出版日期: 2018-05-09
  • 售價: $1,710
  • 貴賓價: 9.5$1,625
  • 語言: 英文
  • 頁數: 423
  • 裝訂: Paperback
  • ISBN: 1788622944
  • ISBN-13: 9781788622943
  • 相關分類: ReinforcementDeepLearning
  • 下單後立即進貨 (約3~4週)

商品描述

Key Features

  • Learn how to deal with the most-common reinforcement learning problems with the best explained practical approach.
  • Fast paced guide to have a better understanding to know everything about RL concepts, framewords, algorithms and many more.
  • Deep dive and learn how to use popular MDPtoolbox package to its maximum extend.

Book Description

Reinforcement learning(RL) allows machines and software agents to act smart and automatically detect the ideal behavior within a specific surrounding, to maximize its performance and productivity. Reinforcement learning is becoming popular and is used as a tool for constructing autonomous systems that improve themselves with experience.

This book will give you a rundown on a brief introduction to reinforcement learning, using popular MDPtoolbox package. We will break the RL framework into its core building blocks, and provide you with details of each of the elements. In this journey you will see, common RL problems like Multi-Armed Bandit problem, types of RL learning algorithms, Markov Decision Processes (MDPs), monte carlo, dynamic programming such as policy and value iteration. Next you will identify temporal difference learnings such as Q-learning and SARSA. You will then learn, that, the utilization of various algorithms in each of these building blocks is kept secondary, as this research area is still open to better algorithms. We will take a practical and simple approach towards explaining the various building blocks of RL, and then bring them together to create a solution.

By the end of this book you will be able to write his/her own codes to construct self-learning autonomous systems. You will finally see, how reinforcement learning plays a big role in computer oriented games such as chess or tic-tac-toe agent.

What you will learn

  • Explore the framework, elements and framework of RL
  • Find the resources available for building RL frameworks
  • Run RL based algorithms on your own with sample examples provided, followed by customized exercises.
  • How to formulate models for the environment
  • Agent based models, Environment interactions, RL formulation (rewards, states, policy, action), Exploration v/s Exploitation, Decision making, Optimization
  • Most recent libraries and packages in R (on RL elements)
  • How to define and evaluation policies with specific mathematical formulation
  • Devise the value functions in a mathematical formulation, and learn the various methodologies/algorithms for the evaluation of policies
  • How RL is different from other supervised/unsupervised algorithms