HAMLET JESSE MEDINA RUIZ holds the position of Chief Data Scientist at Criteo. He specializes in time series forecasting, machine learning, deep learning, and Generative AI. He actively explores the potential of cutting-edge AI technologies, such as Generative AI across diverse applications. He holds an electronic engineering degree from Universidad Rafael Belloso Chacin in Venezuela, as well as two master's degrees with honors in mathematics and machine learning from the Institut Polytechnique de Paris and Université Paris-Saclay. Additionally, he earned a PhD in physics from Université Paris-Saclay. Hamlet has consistently achieved first place and top ten rankings in global machine learning contests, earning the titles of Kaggle Expert and Numerai Expert for these challenges. Recently, he also earned a MicroMaster's in finance from MIT's Sloan School of Management.

ERNEST CHAN (ERNIE) is the Founder and Chief Scientific Officer of [...] ([...] which offers AI-driven adaptive optimization solutions to the finance industry and beyond. He is also the Founder and Non-executive Chairperson of QTS Capital Management ([...] a quantitative CTA/CPO since 2011. He started his career as a machine learning researcher at IBM's T.J. Watson Research Center's language modeling group, which produced some of the best-known quant fund managers. Ernie is the acclaimed author of three previous books, Quantitative Trading (2nd Edition), Algorithmic Trading, and Machine Trading, all published by Wiley. More about these books and Ernie's workshops on topics in quantitative investing and machine learning can be found at [...] He obtained his PhD in physics from Cornell University and his BS in physics from the University of Toronto.

Preface xv

Acknowledgments xix

About the Authors xxi

Part I Generative AI for Trading and Asset Management: A No-code

Introduction 1

Chapter 1 No-code Generative AI for Basic Quantitative Finance 3

1.1 Retrieving Historical Market Data 4

1.2 Computing Sharpe Ratio 7

1.3 Data Formatting and Analysis 8

1.4 Translating Matlab Codes to Python Codes 11

1.5 Conclusion 16

Chapter 2 No-code Generative AI for Trading Strategies Development 17

2.1 Creating Codes from a Strategy Specification 19

2.2 Summarizing a Trading Strategy Paper and Creating Backtest Codes from It 34

2.3 Searching for a Portfolio Optimization Algorithm Based on Machine Learning 45

2.4 Explore Options Term Structure Arbitrage Strategies 50

2.5 Conclusion 64

2.6 Exercises 66

2A.1 Computing Next-day's Return 67

2A.2 Uploading the Fama-French Factors 68

2A.3 Combining Fama-French Factors with Next-day's Returns 68

Chapter 3 Whirlwind Tour of ML in Asset Management 71

3.1 Unsupervised Learning 72

3.2 Supervised Learning 77

3.3 Deep Reinforcement Learning 99

3.4 Data Engineering 100

3.5 Feature Engineering 102

3.6 Conclusion 106

Part II Deep Generative Models for Trading and Asset Management 107

Chapter 4 Understanding Generative AI 109

4.1 Why Generative Models 110

4.2 Difference with Discriminative Models 110

4.3 How Can We Use Them? 111

4.4 Illustrating Generative Models with ChatGPT 113

4.5 Hybrid Modeling: Combining Generative and Discriminative Models 119

4.6 Taxonomy of Generative Models 123

4.7 Conclusion 124

Chapter 5 Deep Autoregressive Models for Sequence Modeling 125

5.1 Representation Complexity 126

5.2 Representation and Complexity Reduction 127

5.3 A Short Tour of Key Model Families 128

5.4 Model Fitting 155

5.5 Conclusions 157

Chapter 6 Deep Latent Variable Models 159

6.1 Introduction 160

6.2 Latent Variable Models 162

6.3 Examples of Traditional Latent Variable Models 162

6.4 Learning 171

6.5 Variational Autoencoder (VAE) 176

6.6 VAEs for Sequential Data and Time Series 177

6.7 Conclusion 181

Chapter 7 Flow Models 183

7.1 Introduction 183

7.2 Model Training 185

7.3 Linear Flows 185

7.4 Designing Nonlinear Flows 187

7.5 Coupling Flows 188

7.6 Autoregressive Flows 195

7.7 Continuous Normalizing Flows 195

7.8 Modeling Financial Time Series with Flow Models 196

7.9 Conclusion 199

Chapter 8 Generative Adversarial Networks 201

8.1 Introduction 202

8.2 Training 204

8.3 Some Theoretical Insight in GANs 208

8.4 Why Is GAN Training Hard? Improving GAN Training Techniques 209

8.5 Wasserstein GAN (WGAN) 211

8.6 Extending GANs for Time Series 214

8.7 Conclusion 215

Chapter 9 Leveraging LLMs for Sentiment Analysis in Trading 217

9.1 Sentiment Analysis in Fed Press Conference Speeches Using Large Language Models 217

9.2 Data: Video + Market Prices 221

9.3 Speech-to-text Conversion 221

9.4 Sentiment Analysis 225

9.5 Experiment Results 232

9.6 Conclusion 234

Chapter 10 Efficient Inference 235

10.1 Introduction 235

10.2 Scaling Large Language Models: High Performance, High Computational Cost, and Emergent Abilities 236

10.3 Making FinBERT Faster 240

10.4 Model Quantization 247

10.5 Customizing Your LLM: Adapting Models to Your Needs 252

10.6 Conclusions 256

Chapter 11 Afterword 257

11.1 Diffusion Models 260

11.2 Combining Generative Model Variants 260

11.3 LLMs as Financial Advisors 261

References 263

Appendix 271

A.1 Retrieving Adjusted Closing Prices and Computing Daily Returns 271

A.2 Installing Python 273

A.2.1 Step 1: Download Python 273

A.2.2 Step 2: Install Python 274

A.2.3 Step 3: Set Up a Virtual Environment (Optional but Recommended) 274

A.2.4 Step 4: Install Packages with pip 274

A.2.5 Step 5: Consider an Integrated Development Environment (IDE) 274

A.2.6 Additional Tips 275

A.3 Plotting the Risk-free-rate over the Years 276

A.4 Computing the Sharpe Ratio of SPY 278

A.5 Matlab Code for Computing Efficient Frontier and Finding the Tangency Portfolio 280

Index 283