This chapter addresses the curse of dimensionality through value function approximation. Parameterized function approximators represent value functions compactly and generalize across states. The chapter frames approximate policy evaluation as supervised learning. Key theoretical objects include the projection operator and projected Bellman equation. Least squares temporal difference (LSTD) provides a batch solution for linear approximation. Deep Q-networks combining neural networks with experience replay are introduced as the foundation for deep RL successes.

In this chapter, we examine our first supervised learning problem, focusing on how to construct prediction functions and assess their performance. Given data consisting of predictor–response pairs, we can learn the parameters of a prediction function by minimising a loss, such as the residual sum of squares, which measures the discrepancy between actual and predicted responses. Using more flexible families of prediction functions typically reduces loss on the training data, but excessive flexibility can lead to overfitting: fitting to noise rather than the systematic component of the relationship. Overfitting results in poor prediction performance on new, unseen data. To estimate how a prediction method will perform on unseen data, we use cross-validation. However, when we compare many prediction methods using cross-validation, the best-performing method often appears better than it truly is; its apparent performance is an unreliable guide to its future accuracy. Prior knowledge is crucial for selecting plausible prediction methods to compare. Finally, we can use bootstrapping to quantify uncertainty in prediction functions and their predictions.

This chapter provides an overview of the types of inference problems we address and the different approaches to solving them. We focus on risky inference: drawing conclusions, learning and making predictions in situations where certainty is impossible. Predicting a response from one or more predictors using past data is called supervised learning. When the response is continuous, the task is regression; when it is categorical, the task is classification. In unsupervised learning, there is no response variable. Instead, the goal is to find patterns or structure in data, as in density estimation, clustering and dimensionality reduction. In both supervised and unsupervised contexts, overfitting occurs when we model data in excessive detail and fail to distinguish systematic patterns from noise; underfitting occurs when our models are too simple to capture systematic patterns. Probability is a key tool for tackling risky inference, with frequentist and Bayesian interpretations motivating distinct approaches. Finally, large neural networks have proven remarkably effective in both supervised and unsupervised tasks, often avoiding overfitting despite containing billions of parameters.

In this chapter, we have included the leadership style measured as preference for delegation and uncertainty avoidance and organizational climate measured as tension and readiness for change to the set of relations which should fit together to meet a firm’s goals. There are four leader styles: maestro, manager, leader, and producer; and there are four climates: group, internal process, developmental, and rational goal. We discuss fit and misfit possibilities for all of these. Finally, we show how to make short-term and long-term changes to address misfit situations.

This chapter provides a comprehensive introduction to supervised learning techniques for classification problems. It begins with logistic regression for binary classification, explaining the sigmoid function and gradient ascent optimization. The chapter then covers softmax regression for multi-class problems, followed by k-nearest neighbors (kNN) as an intuitive distance-based classifier.

Decision trees are explored in detail, including entropy, information gain, and the ID3 algorithm, along with derived decision rules and association rules. Random forests are presented as an ensemble method that addresses overfitting by combining multiple decision trees.

The chapter covers Naive Bayes classification based on Bayes’ theorem, despite its "naive" independence assumption. Finally, Support Vector Machines (SVMs) are introduced for both linear and non-linear classification using maximum margin hyperplanes.

Each technique includes hands-on R programming examples with real datasets, practical applications, and exercises to reinforce learning concepts.

This chapter introduces machine learning as a subset of artificial intelligence that enables computers to learn from data and make predictions without explicit programming. It defines machine learning through Tom Mitchell’s formal framework and explores real-world applications like self-driving cars, optical character recognition, and recommendation systems. The chapter focuses on regression as a fundamental machine learning technique, covering both linear modeling approaches and gradient descent algorithms for parameter optimization. Through hands-on examples using R, students learn to implement linear regression and gradient descent from scratch, understanding how models minimize error functions to find optimal parameters. The chapter emphasizes practical application over theoretical derivations.

The chapter introduces fundamental principles of deep learning. We discuss supervised learning of feedforward neural networks by considering a binary classification problem. Gradient descent techniques and backpropagation learning algorithms are introduced as means of training neural networks. The impact of neuron activations and convolutional and residual network architectures on the learning performance are discussed. Finally, regularization techniques such as batch normalization and dropout are introduced for improving the accuracy of trained models. The chapter is essential to connect advances in conventional deep learning algorithms to neuromorphic concepts.

This chapter introduces machine learning as a subset of artificial intelligence that enables computers to learn from data without explicit programming. It defines machine learning using Tom Mitchell’s formal framework and explores practical applications like self-driving cars, optical character recognition, and recommendation systems. The chapter focuses on regression as a fundamental machine learning technique, explaining linear regression for modeling relationships between variables. A key section covers gradient descent, an optimization algorithm that iteratively finds the best model parameters by minimizing error functions. Through hands-on Python examples, students learn to implement both linear regression and gradient descent algorithms, visualizing how models improve over iterations. The chapter emphasizes practical considerations for choosing appropriate algorithms, including accuracy, training time, linearity assumptions, and the number of parameters, preparing students for more advanced supervised and unsupervised learning techniques.

This chapter explores supervised learning techniques where algorithms learn from labeled training data to make predictions. It begins with logistic regression for binary classification problems, using the sigmoid function to output probabilities between 0 and 1. Softmax regression extends this to multi-class problems. The chapter covers k-nearest neighbors (kNN), which classifies data points based on their similarity to training examples. Decision trees use entropy and information gain to create interpretable classification rules, while random forests combine multiple decision trees to reduce overfitting through ensemble methods. Naive Bayes applies Bayes’ theorem with independence assumptions for probabilistic classification, particularly effective for text classification. Finally, support vector machines (SVM) find optimal decision boundaries by maximizing margins between classes. Each technique is demonstrated through hands-on Python examples using real datasets, showing practical applications in various domains from healthcare to finance.

This chapter educates the reader on the main ideas that have enabled various advancements in Artificial Intelligence (AI) and Machine Learning (ML). Using various examples, and taking the reader on a journey through history, it showcases how the main ideas developed by the pioneers of AI and ML are being used in our modern era to make the world a better place. It communicates that our lives are surrounded by algorithms that work based on a few main ideas. It also discusses recent advancements in Generative AI, including the main ideas that led to the creation of Large Language Models (LLMs) such as Chat GPT. The chapter also discusses various societal considerations in AI and ML and ends with various technological advancements that could further improve our abilities in using the main ideas.

An introduction to AI, including an overview of essential technologies such as machine learning and deep learning, and a discussion on generative AI and its potential limitations. The chapter includes an exploration of AI's history, including its relationship to cybernetics, its role as a codebreaker, periods of optimism and “AI winters,” and today's global development with generative AI. Chapter 1 also include an analysis of AI's role in the international and national context, focusing on potential conflicts of goals and threats that can arise from technology.

Generative artificial intelligence has a long history but surged into global prominence with the introduction in 2017 of the transformer architecture for large language models. Based on deep learning with artificial neural networks, transformers revolutionised the field of generative AI for production of natural language outputs. Today’s large language models, and other forms of generative artificial intelligence, now have unprecedented capability and versatility. This emergence of these forms of highly capable generative AI poses many legal issues and questions, including consequences for intellectual property, contracts and licences, liability, data protection, use in specific sectors, potential harms, and of course ethics, policy, and regulation of the technology. To support the discussion of these topics in this Handbook, this chapter gives a relatively non-technical introduction to the technology of modern artificial intelligence and generative AI.

In this chapter, we introduce the reader to basic concepts in machine learning. We start by defining the artificial intelligence, machine learning, and deep learning. We give a historical viewpoint on the field, also from the perspective of statistical physics. Then, we give a very basic introduction to different tasks that are amenable for machine learning such as regression or classification and explain various types of learning. We end the chapter by explaining how to read the book and how chapters depend on each other.

ML methods are increasingly being used in (corporate) finance studies, with impressive applications. ML methods can be applied with the aim of reducing prediction error in the models, but can also be used to extend the existing traditional econometric methods. The performance of the ML models depends on the quality of the input data and the choice of model. There are many ML models, but all come with their own specific details. It is therefore essential to select accurate model(s) for the analysis. This chapter briefly reviews some broad types of ML methods. It covers supervised learning, which tends to achieve superior prediction performance by using more flexible functional forms than OLS in the prediction model. It explains unsupervised learning methods that derive and learn structural information from conventional data. Finally, the chapter also discusses some limitations and drawbacks of ML, as well as potential remedies.