Programming Machine Learning

Table of contents

1.  Acknowledgments
2.  How the Heck Is That Possible?
   1. About This Book
   2. Before We Begin
3. Part I. From Zero to Image Recognition
   1. 1. How Machine Learning Works
      1. Programming vs. Machine Learning
      2. Supervised Learning
      3. The Math Behind the Magic
      4. Setting Up Your System
   2. 2. Your First Learning Program
      1. Getting to Know the Problem
      2. Coding Linear Regression
      3. Adding a Bias
      4. What You Just Learned
      5. Hands On: Tweaking the Learning Rate
   3. 3. Walking the Gradient
      1. Our Algorithm Doesn’t Cut It
      2. Gradient Descent
      3. What You Just Learned
      4. Hands On: Basecamp Overshooting
   4. 4. Hyperspace!
      1. Adding More Dimensions
      2. Matrix Math
      3. Upgrading the Learner
      4. Bye Bye, Bias
      5. A Final Test Drive
      6. What You Just Learned
      7. Hands On: Field Statistician
   5. 5. A Discerning Machine
      1. Where Linear Regression Fails
      2. Invasion of the Sigmoids
      3. Classification in Action
      4. What You Just Learned
      5. Hands On: Weighty Decisions
   6. 6. Getting Real
      1. Data Come First
      2. Our Own MNIST Library
      3. The Real Thing
      4. What You Just Learned
      5. Hands On: Tricky Digits
   7. 7. The Final Challenge
      1. Going Multiclass
      2. Moment of Truth
      3. What You Just Learned
      4. Hands On: Minesweeper
   8. 8. The Perceptron
      1. Enter the Perceptron
      2. Assembling Perceptrons
      3. Where Perceptrons Fail
      4. A Tale of Perceptrons
4. Part II. Neural Networks
   1. 9. Designing the Network
      1. Assembling a Neural Network from Perceptrons
      2. Enter the Softmax
      3. Here’s the Plan
      4. What You Just Learned
      5. Hands On: Network Adventures
   2. 10. Building the Network
      1. Coding Forward Propagation
      2. Cross Entropy
      3. What You Just Learned
      4. Hands On: Time Travel Testing
   3. 11. Training the Network
      1. The Case for Backpropagation
      2. From the Chain Rule to Backpropagation
      3. Applying Backpropagation
      4. Initializing the Weights
      5. The Finished Network
      6. What You Just Learned
      7. Hands On: Starting Off Wrong
   4. 12. How Classifiers Work
      1. Tracing a Boundary
      2. Bending the Boundary
      3. What You Just Learned
      4. Hands On: Data from Hell
   5. 13. Batchin’ Up
      1. Learning, Visualized
      2. Batch by Batch
      3. Understanding Batches
      4. What You Just Learned
      5. Hands On: The Smallest Batch
   6. 14. The Zen of Testing
      1. The Threat of Overfitting
      2. A Testing Conundrum
      3. What You Just Learned
      4. Hands On: Thinking About Testing
   7. 15. Let’s Do Development
      1. Preparing Data
      2. Tuning Hyperparameters
      3. The Final Test
      4. Hands On: Achieving 99%
      5. What You Just Learned… and the Road Ahead
5. Part III. Deep Learning
   1. 16. A Deeper Kind of Network
      1. The Echidna Dataset
      2. Building a Neural Network with Keras
      3. Making It Deep
      4. What You Just Learned
      5. Hands On: Keras Playground
   2. 17. Defeating Overfitting
      1. Overfitting Explained
      2. Regularizing the Model
      3. A Regularization Toolbox
      4. What You Just Learned
      5. Hands On: Keeping It Simple
   3. 18. Taming Deep Networks
      1. Understanding Activation Functions
      2. Beyond the Sigmoid
      3. Adding More Tricks to Your Bag
      4. What You Just Learned
      5. Hands On: The 10 Epochs Challenge
   4. 19. Beyond Vanilla Networks
      1. The CIFAR-10 Dataset
      2. The Building Blocks of CNNs
      3. Running on Convolutions
      4. What You Just Learned
      5. Hands On: Hyperparameters Galore
   5. 20. Into the Deep
      1. The Rise of Deep Learning
      2. Unreasonable Effectiveness
      3. Where Now?
      4. Your Journey Begins
6. A1. Just Enough Python
   1. What Python Looks Like
   2. Python’s Building Blocks
   3. Defining and Calling Functions
   4. Working with Modules and Packages
   5. Creating and Using Objects
   6. That’s It, Folks!
7. A2. The Words of Machine Learning