Applying Reinforcement Learning on Real-World Data with Practical Examples in Python

Cover
Copyright Page
Title Page
Contents
Preface
Acknowledgments
Background and Definitions
	What is Real-World Reinforcement Learning?
	Reinforcement Learning vs. Other Machine Learning Techniques
	What About Deep Learning?
	Notable Real-World Examples
	Supporting Code
		What Form is the Code Published in?
	Summary of Learning Procedure
		The Markov Decision Process Environment
		The Learning Components
		Model-Based vs. Model-Free Approaches
		Returns and Episodes
Reinforcement Learning Theory
	Core Principles of Reinforcement Learning
	Model-Based Methods
		Policy Evaluation
		Policy Improvement
		Value Iteration
	Model-Free Methods
		Q-Learning: Off-Policy TD Control
		SARSA: On-Policy TD Control
		Temporal Difference
	Online vs. Offline Reinforcement Learning
A Robot Cleaner Example
	The Task
	Applying Dynamic Programming Methods
		Implementing Value Iteration
		Value Iteration Update Example
	Applying Model-Free Methods
		Implementing the Q-Learning Algorithm
		Understanding the Agent\'s Parameters
		What Remains Unaddressed?
	Understanding the Environment Generation
		The Pre-Built Markov Decision Process
		The Robot\'s Sensory Data
		Defining the MDP
	Conclusion of the Robot Cleaner Example
The Classroom Environment
	Description of the Real-World Problem
		The Agent\'s Role in This Problem
		Defining the Reward Signal
		Defining Episodes
		Exploration vs. Exploitation
	Environmental Definition
		Mimicking the Real-World with a Simulation
	Data Collection Methods
		An Online Setup
		A Batch-Offline Setup
	Conclusion of the Classroom Example
Industry Applications
	Google Data Center
	Salesforce Text Summarization
	IBM Trading
	PepsiCo Cheetos
	Siam Cement Group Public Company Limited
Conclusion
Bibliography
Authors\' Biographies