This book is dedicated to those who tirelessly contribute to advancing the field of machine learning through research and development. ## Your passion for discovery and innovation and your commitment to sharing knowledge and resources through the open-source community is an inspiration to us all. 1I will use machine learning as an umbrella term for machine learning, deep learning, and artificial intelligence. ## Who Is This Book For? 2 ## Who Is This Book For? This book is for people with a beginner or intermediate background in machine learning who want to learn something new. This book will expose you to new concepts and ideas if you are already familiar with machine learning. However, it is not a math or coding book. You won't need to solve any proofs or run any code while reading. In other words, this book is a perfect travel companion or something you can read on your favorite reading chair with your morning coffee. Table of Contents 4 Preface 6 Who Is This Book For? 7 What Will You Get Out of This Book? 8 How To Read This Book 9 Discussion Forum 11 Sharing Feedback and Supporting This Book 12 Acknowledgements 13 About the Author 14 Copyright and Disclaimer 15 Credits 16 Introduction 17 Chapter 1. Neural Networks and Deep Learning 18 Q1. Embeddings, Representations, and Latent Space 19 Q2. Self-Supervised Learning 23 Q3. Few-Shot Learning 30 Q4. The Lottery Ticket Hypothesis 34 Q5. Reducing Overfitting with Data 37 Q6. Reducing Overfitting with Model Modifications 42 Q7. Multi-GPU Training Paradigms 50 Q8. The Keys to Success of Transformers 57 Q9. Generative AI Models 62 Q10. Sources of Randomness 75 Chapter 2. Computer Vision 84 Q11. Calculating the Number of Parameters 85 Q12. The Equivalence of Fully Connected and Convolutional Layers 90 Q13. Large Training Sets for Vision Transformers 94 Chapter 3. Natural Language Processing 103 Q15. The Distributional Hypothesis 104 Q16. Data Augmentation for Text 109 Q17. ``Self''-Attention 116 Q18. Encoder- And Decoder-Style Transformers 121 Q19. Using and Finetuning Pretrained Transformers 131 Q20. Evaluating Generative Language Models 146 Chapter 4. Production, Real-World, And Deployment Scenarios 157 Q21. Stateless And Stateful Training 158 Q22. Data-Centric AI 161 Q23. Speeding Up Inference 165 Chapter 5. Predictive Performance and Model Evaluation 172 Q25. Poisson and Ordinal Regression 173 Q27. Proper Metrics 175 Q28. The k in k-fold cross-validation 180 Q29. Training and Test Set Discordance 184 Q30. Limited Labeled Data 187 Afterword 201 Appendix A: Reader Quiz Solutions 202 Appendix B: List of Questions 228
