You're reading from Machine Learning with PyTorch and Scikit-Learn Develop machine learning and deep learning models with Python

Product type Paperback

Published in Feb 2022

Publisher Packt

ISBN-13 9781801819312

Length 774 pages

Edition 1st Edition

Languages

Python

Tools

PyTorch

Concepts

Deep Learning

Authors (3):

Sebastian Raschka

Yuxi (Hayden) Liu

Vahid Mirjalili

View More author details

Table of Contents (22) Chapters

Preface

1. Giving Computers the Ability to Learn from Data FREE CHAPTER

2. Training Simple Machine Learning Algorithms for Classification

3. A Tour of Machine Learning Classifiers Using Scikit-Learn

4. Building Good Training Datasets – Data Preprocessing

5. Compressing Data via Dimensionality Reduction

6. Learning Best Practices for Model Evaluation and Hyperparameter Tuning

7. Combining Different Models for Ensemble Learning

8. Applying Machine Learning to Sentiment Analysis

9. Predicting Continuous Target Variables with Regression Analysis

10. Working with Unlabeled Data – Clustering Analysis

11. Implementing a Multilayer Artificial Neural Network from Scratch

12. Parallelizing Neural Network Training with PyTorch

13. Going Deeper – The Mechanics of PyTorch

14. Classifying Images with Deep Convolutional Neural Networks

15. Modeling Sequential Data Using Recurrent Neural Networks

16. Transformers – Improving Natural Language Processing with Attention Mechanisms

17. Generative Adversarial Networks for Synthesizing New Data

18. Graph Neural Networks for Capturing Dependencies in Graph Structured Data

19. Reinforcement Learning for Decision Making in Complex Environments

20. Other Books You May Enjoy

21. Index

A glance at deep Q-learning

In the previous code, we saw an implementation of the popular Q-learning algorithm for the grid world example. This example consisted of a discrete state space of size 30, where it was sufficient to store the Q-values in a Python dictionary.

However, we should note that sometimes the number of states can get very large, possibly almost infinitely large. Also, we may be dealing with a continuous state space instead of working with discrete states. Moreover, some states may not be visited at all during training, which can be problematic when generalizing the agent to deal with such unseen states later.

To address these problems, instead of representing the value function in a tabular format like V(S_t), or Q(S_t, A_t), for the action-value function, we use a function approximation approach. Here, we define a parametric function, v_w(x_s), that can learn to approximate the true value function, that is, , where x_s is a set of input features (or “...