Reinforcement Learning with Python Explained for Beginners

Reinforcement Learning with Python Explained for Beginners

by AI Sciences OU
Released February 2021
Publisher(s): Packt Publishing
ISBN: 9781801072274

Watch it now on the O’Reilly learning platform with a 10-day free trial.

O’Reilly members get unlimited access to books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.

Start your free trial

Video description

Although introduced academically decades ago, the recent developments in the field of reinforcement learning have been phenomenal. Domains such as self-driving cars, natural language processing, healthcare industry, online recommender systems, and so on have already seen how RL-based AI agents can bring tremendous gains.

This course will help you get started with reinforcement learning first by establishing the motivation for this field and then covering all the essential topics, such as Markov Decision Processes, policy and rewards, model-free learning, temporal difference learning, and so on.

Each topic is accompanied by exercises and complementing analysis to help you gain practical and tangible coding skills.

By the end of this course, not only will you have gained the necessary understanding to implement RL in your projects but also implemented an actual Frozenlake project using the OpenAI Gym toolkit.

What You Will Learn

  • Understand the motivation for reinforcement learning
  • Understand all the elements of a Markov Decision Process
  • Learn how to model uncertainty of the environments
  • Solve Markov Decision Processes
  • Implement temporal difference learning and Q-learning in Python
  • Execute the Frozenlake project using the OpenAI Gym toolkit

Audience

This course is designed for beginners in the field of data science and machine learning. Anyone who wants to learn RL and apply it in realistic projects would benefit from this course.

About The Author

AI Sciences: AI Sciences are experts, PhDs, and artificial intelligence practitioners, including computer science, machine learning, and Statistics. Some work in big companies such as Amazon, Google, Facebook, Microsoft, KPMG, BCG, and IBM.

AI sciences produce a series of courses dedicated to beginners and newcomers on techniques and methods of machine learning, statistics, artificial intelligence, and data science. They aim to help those who wish to understand techniques more easily and start with less theory and less extended reading. Today, they publish more comprehensive courses on specific topics for wider audiences.

Their courses have successfully helped more than 100,000 students master AI and data science.

Table of contents

  1. Chapter 1 : Introduction to Course and Instructor
    1. Introduction to Course and Instructor
  2. Chapter 2 : Motivation Reinforcement Learning
    1. What is Reinforcement Learning
    2. What is Reinforcement Learning Hiders and Seekers by OpenAI
    3. RL Versus Other ML Frameworks
    4. Why Reinforcement Learning
    5. Examples of Reinforcement Learning
    6. Limitations of Reinforcement Learning
    7. Exercises
  3. Chapter 3 : Terminology of Reinforcement Learning
    1. What is Environment
    2. What is Environment_2
    3. What is Agent
    4. What is State
    5. State Belongs to Environment and not to Agent
    6. What is Action
    7. What is Reward
    8. Goal
    9. Policy
    10. Summary
  4. Chapter 4 : GridWorld Example
    1. Setup 1
    2. Setup 2
    3. Setup 3
    4. Policy Comparison
    5. Deterministic Environment
    6. Stochastic Environment
    7. Stochastic Environment 2
    8. Stochastic Environment 3
    9. Non-Stationary Environment
    10. GridWorld Summary
    11. Activity
  5. Chapter 5 : Markov Decision Process Prerequisites
    1. Probability
    2. Probability 2
    3. Probability 3
    4. Conditional Probability
    5. Conditional Probability Fun Example
    6. Joint Probability
    7. Joint probability 2
    8. Joint probability 3
    9. Expected Value
    10. Conditional Expectation
    11. Modeling Uncertainty of Environment
    12. Modeling Uncertainty of Environment 2
    13. Modeling Uncertainty of Environment 3
    14. Modeling Uncertainty of Environment Stochastic Policy
    15. Modeling Uncertainty of Environment Stochastic Policy 2
    16. Modeling Uncertainty of Environment Value Functions
    17. Running Averages
    18. Running Averages 2
    19. Running Averages as Temporal Difference
    20. Activity
  6. Chapter 6 : Elements of Markov Decision Process
    1. Markov Property
    2. State Space
    3. Action Space
    4. Transition Probabilities
    5. Reward Function
    6. Discount Factor
    7. Summary
    8. Activity
  7. Chapter 7 : More on Reward
    1. MOR Quiz 1
    2. MOR Quiz Solution 1
    3. MOR Quiz 2
    4. MOR Quiz Solution 2
    5. MOR Reward Scaling
    6. MOR Infinite Horizons
    7. MOR Quiz 3
    8. MOR Quiz Solution 3
  8. Chapter 8 : Solving Markov DP
    1. MDP Recap
    2. Value Functions
    3. Optimal Value Function
    4. Optimal Policy
    5. Bellman Equation
    6. Value Iteration
    7. Value Iteration Quiz
    8. Value Iteration Quiz Gamma Missing
    9. Value Iteration Solution
    10. Problems of Value Iteration
    11. Policy Evaluation
    12. Policy Evaluation 2
    13. Policy Evaluation 3
    14. Policy Evaluation d Form Solution
    15. Policy Iteration
    16. State Action Values
    17. V and Q Comparisons
  9. Chapter 9 : Value Approximation
    1. What Does it Mean that MDP is Unknown
    2. Why Transition Probabilities are Important
    3. Model-Based Solutions
    4. Model-Free Solutions
    5. Monte-Carlo Learning
    6. Monte-Carlo Learning Example
    7. Monte-Carlo Learning Limitations
  10. Chapter 10 : Temporal Differencing - Q Learning
    1. Running Average
    2. Learning Rate
    3. Learning Equation
    4. TD Algorithm
    5. Exploration Versus Exploitation
    6. Epsilon Greedy Policy
    7. SARSA
    8. Q-Learning
    9. Q-Learning Implementation for MAPROVER Clipped
  11. Chapter 11 : TD Lambda
    1. N-Step Look a Head
    2. Formulation
    3. Values
    4. TD Q-Learning TD Lambda
    5. TD Q-Learning TD Lambda TD(Lambda) MAPRover Activity
  12. Chapter 12 : Project Frozenlake (Open AI Gym)
    1. Frozenlake 1
    2. Frozenlake Implementation

Product information

  • Title: Reinforcement Learning with Python Explained for Beginners
  • Author(s): AI Sciences OU
  • Release date: February 2021
  • Publisher(s): Packt Publishing
  • ISBN: 9781801072274