Machine Learning: Foundations, Methodologies, and Applications – serie

This book introduces readers to an evolutionary learning approach, specifically genetic programming (GP), for production scheduling. The book is divided into six parts. In Part I, it provides an introduction to production scheduling, existing solution methods, and the GP approach to production scheduling. Characteristics of production environments, problem formulations, an abstract GP framework for production scheduling, and evaluation criteria are also presented. Part II shows various ways that GP can be employed to solve static production scheduling problems and their connections with conventional operation research methods. In turn, Part III shows how to design GP algorithms for dynamic production scheduling problems and describes advanced techniques for enhancing GP’s performance, including feature selection, surrogate modeling, and specialized genetic operators. In Part IV, the book addresses how to use heuristics to deal with multiple, potentially conflicting objectives in production scheduling problems, and presents an advanced multi-objective approach with cooperative coevolution techniques or multi-tree representations. Part V demonstrates how to use multitask learning techniques in the hyper-heuristics space for production scheduling. It also shows how surrogate techniques and assisted task selection strategies can benefit multitask learning with GP for learning heuristics in the context of production scheduling. Part VI rounds out the text with an outlook on the future.Given its scope, the book benefits scientists, engineers, researchers, practitioners, postgraduates, and undergraduates in the areas of machine learning, artificial intelligence, evolutionary computation, operations research, and industrial engineering.

This book introduces readers to an evolutionary learning approach, specifically genetic programming (GP), for production scheduling. The book is divided into six parts. In Part I, it provides an introduction to production scheduling, existing solution methods, and the GP approach to production scheduling. Characteristics of production environments, problem formulations, an abstract GP framework for production scheduling, and evaluation criteria are also presented. Part II shows various ways that GP can be employed to solve static production scheduling problems and their connections with conventional operation research methods. In turn, Part III shows how to design GP algorithms for dynamic production scheduling problems and describes advanced techniques for enhancing GP’s performance, including feature selection, surrogate modeling, and specialized genetic operators. In Part IV, the book addresses how to use heuristics to deal with multiple, potentially conflicting objectives in production scheduling problems, and presents an advanced multi-objective approach with cooperative coevolution techniques or multi-tree representations. Part V demonstrates how to use multitask learning techniques in the hyper-heuristics space for production scheduling. It also shows how surrogate techniques and assisted task selection strategies can benefit multitask learning with GP for learning heuristics in the context of production scheduling. Part VI rounds out the text with an outlook on the future.Given its scope, the book benefits scientists, engineers, researchers, practitioners, postgraduates, and undergraduates in the areas of machine learning, artificial intelligence, evolutionary computation, operations research, and industrial engineering.

Machine learning (ML) has become a commonplace element in our everyday lives and a standard tool for many fields of science and engineering. To make optimal use of ML, it is essential to understand its underlying principles. This book approaches ML as the computational implementation of the scientific principle. This principle consists of continuously adapting a model of a given data-generating phenomenon by minimizing some form of loss incurred by its predictions. The book trains readers to break down various ML applications and methods in terms of data, model, and loss, thus helping them to choose from the vast range of ready-made ML methods.The book’s three-component approach to ML provides uniform coverage of a wide range of concepts and techniques. As a case in point, techniques for regularization, privacy-preservation as well as explainability amount tospecific design choices for the model, data, and loss of a ML method.

Machine learning (ML) has become a commonplace element in our everyday lives and a standard tool for many fields of science and engineering. To make optimal use of ML, it is essential to understand its underlying principles. This book approaches ML as the computational implementation of the scientific principle. This principle consists of continuously adapting a model of a given data-generating phenomenon by minimizing some form of loss incurred by its predictions. The book trains readers to break down various ML applications and methods in terms of data, model, and loss, thus helping them to choose from the vast range of ready-made ML methods.The book’s three-component approach to ML provides uniform coverage of a wide range of concepts and techniques. As a case in point, techniques for regularization, privacy-preservation as well as explainability amount tospecific design choices for the model, data, and loss of a ML method.

Deep learning has significantly reshaped a variety of technologies, such as image processing, natural language processing, and audio processing. The excellent generalizability of deep learning is like a “cloud” to conventional complexity-based learning theory: the over-parameterization of deep learning makes almost all existing tools vacuous. This irreconciliation considerably undermines the confidence of deploying deep learning to security-critical areas, including autonomous vehicles and medical diagnosis, where small algorithmic mistakes can lead to fatal disasters. This book seeks to explaining the excellent generalizability, including generalization analysis via the size-independent complexity measures, the role of optimization in understanding the generalizability, and the relationship between generalizability and ethical/security issues. The efforts to understand the excellent generalizability are following two major paths: (1) developing size-independent complexity measures, which can evaluate the “effective” hypothesis complexity that can be learned, instead of the whole hypothesis space; and (2) modelling the learned hypothesis through stochastic gradient methods, the dominant optimizers in deep learning, via stochastic differential functions and the geometry of the associated loss functions. Related works discover that over-parameterization surprisingly bring many good properties to the loss functions. Rising concerns of deep learning are seen on the ethical and security issues, including privacy preservation and adversarial robustness. Related works also reveal an interplay between them and generalizability: a good generalizability usually means a good privacy-preserving ability; and more robust algorithms might have a worse generalizability.We expect readers can have a big picture of the current knowledge in deep learning theory, understand how the deep learning theory can guide new algorithm designing, and identify future research directions. Readers need knowledge of calculus, linear algebra, probability, statistics, and statistical learning theory.

Entering the field of artificial intelligence and data science can seem daunting to beginners with little to no prior background, especially those with no programming experience. The concepts used in self-driving cars and virtual assistants like Amazon’s Alexa may seem very complex and difficult to grasp. The aim of Artificial Intelligence in Python is to make AI accessible and easy to understand for people with little to no programming experience though practical exercises. Newcomers will gain the necessary knowledge on how to create such systems, which are capable of executing tasks that require some form of human-like intelligence.This book introduces readers to various topics and examples of programming in Python, as well as key concepts in artificial intelligence. Python programming skills will be imparted as we go along. Concepts and code snippets will be covered in a step-by-step manner, to guide and instill confidence in beginners. Complex subjectsin deep learning and machine learning will be broken down into easy-to-digest content and examples. Artificial intelligence implementations will also be shared, allowing beginners to generate their own artificial intelligence algorithms for reinforcement learning, style transfer, chatbots, speech, and natural language processing.

A remarkable facet of the human brain is its ability to manage multiple tasks with apparent simultaneity. Knowledge learned from one task can then be used to enhance problem-solving in other related tasks. In machine learning, the idea of leveraging relevant information across related tasks as inductive biases to enhance learning performance has attracted significant interest. In contrast, attempts to emulate the human brain’s ability to generalize in optimization – particularly in population-based evolutionary algorithms – have received little attention to date.  Recently, a novel evolutionary search paradigm, Evolutionary Multi-Task (EMT) optimization, has been proposed in the realm of evolutionary computation. In contrast to traditional evolutionary searches, which solve a single task in a single run, evolutionary multi-tasking algorithm conducts searches concurrently on multiple search spaces corresponding to different tasks or optimization problems,each possessing a unique function landscape. By exploiting the latent synergies among distinct problems, the superior search performance of EMT optimization in terms of solution quality and convergence speed has been demonstrated in a variety of continuous, discrete, and hybrid (mixture of continuous and discrete) tasks.  This book discusses the foundations and methodologies of developing evolutionary multi-tasking algorithms for complex optimization, including in domains characterized by factors such as multiple objectives of interest, high-dimensional search spaces and NP-hardness.

A remarkable facet of the human brain is its ability to manage multiple tasks with apparent simultaneity. Knowledge learned from one task can then be used to enhance problem-solving in other related tasks. In machine learning, the idea of leveraging relevant information across related tasks as inductive biases to enhance learning performance has attracted significant interest. In contrast, attempts to emulate the human brain’s ability to generalize in optimization – particularly in population-based evolutionary algorithms – have received little attention to date.  Recently, a novel evolutionary search paradigm, Evolutionary Multi-Task (EMT) optimization, has been proposed in the realm of evolutionary computation. In contrast to traditional evolutionary searches, which solve a single task in a single run, evolutionary multi-tasking algorithm conducts searches concurrently on multiple search spaces corresponding to different tasks or optimization problems,each possessing a unique function landscape. By exploiting the latent synergies among distinct problems, the superior search performance of EMT optimization in terms of solution quality and convergence speed has been demonstrated in a variety of continuous, discrete, and hybrid (mixture of continuous and discrete) tasks.  This book discusses the foundations and methodologies of developing evolutionary multi-tasking algorithms for complex optimization, including in domains characterized by factors such as multiple objectives of interest, high-dimensional search spaces and NP-hardness.

This book offers a brief but effective introduction to quantum machine learning (QML). QML is not merely a translation of classical machine learning techniques into the language of quantum computing, but rather a new approach to data representation and processing. Accordingly, the content is not divided into a “classical part” that describes standard machine learning schemes and a “quantum part” that addresses their quantum counterparts. Instead, to immerse the reader in the quantum realm from the outset, the book starts from fundamental notions of quantum mechanics and quantum computing. Avoiding unnecessary details, it presents the concepts and mathematical tools that are essential for the required quantum formalism. In turn, it reviews those quantum algorithms most relevant to machine learning. Later chapters highlight the latest advances in this field and discuss the most promising directions for future research.To gain the most from this book, a basic grasp of statistics and linear algebra is sufficient; no previous experience with quantum computing or machine learning is needed. The book is aimed at researchers and students with no background in quantum physics and is also suitable for physicists looking to enter the field of QML.

This book offers a brief but effective introduction to quantum machine learning (QML). QML is not merely a translation of classical machine learning techniques into the language of quantum computing, but rather a new approach to data representation and processing. Accordingly, the content is not divided into a “classical part” that describes standard machine learning schemes and a “quantum part” that addresses their quantum counterparts. Instead, to immerse the reader in the quantum realm from the outset, the book starts from fundamental notions of quantum mechanics and quantum computing. Avoiding unnecessary details, it presents the concepts and mathematical tools that are essential for the required quantum formalism. In turn, it reviews those quantum algorithms most relevant to machine learning. Later chapters highlight the latest advances in this field and discuss the most promising directions for future research.To gain the most from this book, a basic grasp of statistics and linear algebra is sufficient; no previous experience with quantum computing or machine learning is needed. The book is aimed at researchers and students with no background in quantum physics and is also suitable for physicists looking to enter the field of QML.

This book introduces readers to the fundamentals of and recent advances in federated learning, focusing on reducing communication costs, improving computational efficiency, and enhancing the security level. Federated learning is a distributed machine learning paradigm which enables model training on a large body of decentralized data. Its goal is to make full use of data across organizations or devices while meeting regulatory, privacy, and security requirements. The book starts with a self-contained introduction to artificial neural networks, deep learning models, supervised learning algorithms, evolutionary algorithms, and evolutionary learning. Concise information is then presented on multi-party secure computation, differential privacy, and homomorphic encryption, followed by a detailed description of federated learning. In turn, the book addresses the latest advances in federate learning research, especially from the perspectives of communication efficiency, evolutionarylearning, and privacy preservation.The book is particularly well suited for graduate students, academic researchers, and industrial practitioners in the field of machine learning and artificial intelligence. It can also be used as a self-learning resource for readers with a science or engineering background, or as a reference text for graduate courses.

This book introduces readers to the fundamentals of and recent advances in federated learning, focusing on reducing communication costs, improving computational efficiency, and enhancing the security level. Federated learning is a distributed machine learning paradigm which enables model training on a large body of decentralized data. Its goal is to make full use of data across organizations or devices while meeting regulatory, privacy, and security requirements. The book starts with a self-contained introduction to artificial neural networks, deep learning models, supervised learning algorithms, evolutionary algorithms, and evolutionary learning. Concise information is then presented on multi-party secure computation, differential privacy, and homomorphic encryption, followed by a detailed description of federated learning. In turn, the book addresses the latest advances in federate learning research, especially from the perspectives of communication efficiency, evolutionarylearning, and privacy preservation.The book is particularly well suited for graduate students, academic researchers, and industrial practitioners in the field of machine learning and artificial intelligence. It can also be used as a self-learning resource for readers with a science or engineering background, or as a reference text for graduate courses.

Transfer learning is one of the most important technologies in the era of artificial intelligence and deep learning. It seeks to leverage existing knowledge by transferring it to another, new domain. Over the years, a number of relevant topics have attracted the interest of the research and application community: transfer learning, pre-training and fine-tuning, domain adaptation, domain generalization, and meta-learning.This book offers a comprehensive tutorial on an overview of transfer learning, introducing new researchers in this area to both classic and more recent algorithms. Most importantly, it takes a “student’s” perspective to introduce all the concepts, theories, algorithms, and applications, allowing readers to quickly and easily enter this area. Accompanying the book, detailed code implementations are provided to better illustrate the core ideas of several important algorithms, presenting good examples for practice.

Transfer learning is one of the most important technologies in the era of artificial intelligence and deep learning. It seeks to leverage existing knowledge by transferring it to another, new domain. Over the years, a number of relevant topics have attracted the interest of the research and application community: transfer learning, pre-training and fine-tuning, domain adaptation, domain generalization, and meta-learning.This book offers a comprehensive tutorial on an overview of transfer learning, introducing new researchers in this area to both classic and more recent algorithms. Most importantly, it takes a “student’s” perspective to introduce all the concepts, theories, algorithms, and applications, allowing readers to quickly and easily enter this area. Accompanying the book, detailed code implementations are provided to better illustrate the core ideas of several important algorithms, presenting good examples for practice.

Graph Neural Networks (GNNs) have revolutionized the way we learn representations from graph-structured data, becoming a cornerstone for applications in social networks, recommendation systems, biology, and beyond. However, mainstream GNNs rely heavily on message passing, an iterative process of propagating information between connected nodes. While powerful, this method often incurs significant computational costs, making efficient training a growing challenge as graph sizes scale up.This book addresses these challenges by offering a comprehensive exploration of efficient GNN training through the lens of data management. It highlights how innovative techniques, rooted in decades of graph processing research, can optimize the entire training process without compromising performance. By focusing on system-level enhancements and practical solutions, it provides actionable strategies to overcome efficiency bottlenecks in large-scale GNN training.Readers will gain a deeper understanding of the graph data lifecycle in GNN training, with examples that demonstrate how data management techniques can significantly enhance scalability and performance. The book is designed for a broad audience, including students, researchers, and professionals, offering clear explanations and practical insights for anyone looking to master efficient GNN training.

This book introduces the prevailing domains of recommender systems and cross-device federated learning, highlighting the latest research progress and prospects regarding cross-device federated recommendation. As a privacy-oriented distributed computing paradigm, cross-device federated learning enables collaborative intelligence across multiple devices while ensuring the security of local data. In this context, ubiquitous recommendation services emerge as a crucial application of device-side AI, making a deep exploration of federated recommendation systems highly significant.This book is self-contained, and each chapter can be comprehended independently. Overall, the book organizes existing efforts in federated recommendation from three different perspectives. The perspective of learning paradigms includes statistical machine learning, deep learning, reinforcement learning, and meta learning, where each has detailed techniques (e.g., different neural building blocks) to present relevant studies. The perspective of privacy computing covers homomorphic encryption, differential privacy, secure multi-party computing, and malicious attacks. More specific encryption and obfuscation techniques, such as randomized response and secret sharing, are involved. The perspective of federated issues discusses communication optimization and fairness perception, which are widely concerned in the cross-device distributed environment. In the end, potential issues and promising directions for future research are identified point by point.This book is especially suitable for researchers working on the application of recommendation algorithms to the privacy-preserving federated scenario. The target audience includes graduate students, academic researchers, and industrial practitioners who specialize in recommender systems, distributed machine learning, information retrieval, information security, or artificial intelligence.

This book introduces the prevailing domains of recommender systems and cross-device federated learning, highlighting the latest research progress and prospects regarding cross-device federated recommendation. As a privacy-oriented distributed computing paradigm, cross-device federated learning enables collaborative intelligence across multiple devices while ensuring the security of local data. In this context, ubiquitous recommendation services emerge as a crucial application of device-side AI, making a deep exploration of federated recommendation systems highly significant.This book is self-contained, and each chapter can be comprehended independently. Overall, the book organizes existing efforts in federated recommendation from three different perspectives. The perspective of learning paradigms includes statistical machine learning, deep learning, reinforcement learning, and meta learning, where each has detailed techniques (e.g., different neural building blocks) to present relevant studies. The perspective of privacy computing covers homomorphic encryption, differential privacy, secure multi-party computing, and malicious attacks. More specific encryption and obfuscation techniques, such as randomized response and secret sharing, are involved. The perspective of federated issues discusses communication optimization and fairness perception, which are widely concerned in the cross-device distributed environment. In the end, potential issues and promising directions for future research are identified point by point.This book is especially suitable for researchers working on the application of recommendation algorithms to the privacy-preserving federated scenario. The target audience includes graduate students, academic researchers, and industrial practitioners who specialize in recommender systems, distributed machine learning, information retrieval, information security, or artificial intelligence.

In recent years, embodied multi-agent systems, including multi-robots, have emerged as essential solution for demanding tasks such as search and rescue, environmental monitoring, and space exploration. Effective collaboration among these agents is crucial but presents significant challenges due to differences in morphology and capabilities, especially in heterogenous systems. While existing books address collaboration control, perception, and learning, there is a gap in focusing on active perception and interactive learning for embodied multi-agent systems.This book aims to bridge this gap by establishing a unified framework for perception and learning in embodied multi-agent systems. It presents and discusses the perception-action-learning loop, offering systematic solutions for various types of agents—homogeneous, heterogeneous, and ad hoc. Beyond the popular reinforcement learning techniques, the book provides insights into using fundamental models to tackle complex collaboration problems.By interchangeably utilizing constrained optimization, reinforcement learning, and fundamental models, this book offers a comprehensive toolkit for solving different types of embodied multi-agent problems. Readers will gain an understanding of the advantages and disadvantages of each method for various tasks. This book will be particularly valuable to graduate students and professional researchers in robotics and machine learning. It provides a robust learning framework for addressing practical challenges in embodied multi-agent systems and demonstrates the promising potential of fundamental models for scenario generation, policy learning, and planning in complex collaboration problems.

In recent years, embodied multi-agent systems, including multi-robots, have emerged as essential solution for demanding tasks such as search and rescue, environmental monitoring, and space exploration. Effective collaboration among these agents is crucial but presents significant challenges due to differences in morphology and capabilities, especially in heterogenous systems. While existing books address collaboration control, perception, and learning, there is a gap in focusing on active perception and interactive learning for embodied multi-agent systems.This book aims to bridge this gap by establishing a unified framework for perception and learning in embodied multi-agent systems. It presents and discusses the perception-action-learning loop, offering systematic solutions for various types of agents—homogeneous, heterogeneous, and ad hoc. Beyond the popular reinforcement learning techniques, the book provides insights into using fundamental models to tackle complex collaboration problems.By interchangeably utilizing constrained optimization, reinforcement learning, and fundamental models, this book offers a comprehensive toolkit for solving different types of embodied multi-agent problems. Readers will gain an understanding of the advantages and disadvantages of each method for various tasks. This book will be particularly valuable to graduate students and professional researchers in robotics and machine learning. It provides a robust learning framework for addressing practical challenges in embodied multi-agent systems and demonstrates the promising potential of fundamental models for scenario generation, policy learning, and planning in complex collaboration problems.

This book offers a pioneering exploration of classification-based derivative-free optimization (DFO), providing researchers and professionals in artificial intelligence, machine learning, AutoML, and optimization with a robust framework for addressing complex, large-scale problems where gradients are unavailable. By bridging theoretical foundations with practical implementations, it fills critical gaps in the field, making it an indispensable resource for both academic and industrial audiences.The book introduces innovative frameworks such as sampling-and-classification (SAC) and sampling-and-learning (SAL), which underpin cutting-edge algorithms like Racos and SRacos. These methods are designed to excel in challenging optimization scenarios, including high-dimensional search spaces, noisy environments, and parallel computing. A dedicated section on the ZOOpt toolbox provides practical tools for implementing these algorithms effectively. The book’s structure moves from foundational principles and algorithmic development to advanced topics and real-world applications, such as hyperparameter tuning, neural architecture search, and algorithm selection in AutoML.Readers will benefit from a comprehensive yet concise presentation of modern DFO methods, gaining theoretical insights and practical tools to enhance their research and problem-solving capabilities. A foundational understanding of machine learning, probability theory, and algorithms is recommended for readers to fully engage with the material.

This book offers a pioneering exploration of classification-based derivative-free optimization (DFO), providing researchers and professionals in artificial intelligence, machine learning, AutoML, and optimization with a robust framework for addressing complex, large-scale problems where gradients are unavailable. By bridging theoretical foundations with practical implementations, it fills critical gaps in the field, making it an indispensable resource for both academic and industrial audiences.The book introduces innovative frameworks such as sampling-and-classification (SAC) and sampling-and-learning (SAL), which underpin cutting-edge algorithms like Racos and SRacos. These methods are designed to excel in challenging optimization scenarios, including high-dimensional search spaces, noisy environments, and parallel computing. A dedicated section on the ZOOpt toolbox provides practical tools for implementing these algorithms effectively. The book’s structure moves from foundational principles and algorithmic development to advanced topics and real-world applications, such as hyperparameter tuning, neural architecture search, and algorithm selection in AutoML.Readers will benefit from a comprehensive yet concise presentation of modern DFO methods, gaining theoretical insights and practical tools to enhance their research and problem-solving capabilities. A foundational understanding of machine learning, probability theory, and algorithms is recommended for readers to fully engage with the material.

As a well-known text mining tool, topic modeling can effectively discover the latent semantic structure of text data. Extracting topics from documents is also one of the fundamental challenges in natural language processing. Although topic models have seen significant achievements over the past three decades, there remains a scarcity of methods that effectively model temporal aspect. Moreover, many contemporary topic models continue to grapple with the issue of noise contamination, particularly in social media data.This book presents several approaches designed to address these two limitations. Initially, traditional lifelong topic models aim to accumulate knowledge learned from experience for future task. However, the sequence of topics extracted by these methods may shift over time, leading to semantic misalignment between the topic representations across document streams. Such misalignment can degrade the performances of various downstream tasks, including online document classification and dynamic information retrieval at the topic level. Additionally, the challenge of coherent topic modeling is particularly relevant due to the noise and large scale of social media datasets. Messages on social media platforms often consists of only a few words, resulting in a lack of significant context. Models applied directly to this type of text frequently encounter the problem of feature sparsity, which can yield unsatisfactory outcomes.In the context of emotion detection, public emotions are known to fluctuate across different topics, and topics can evoke public emotion. Thus, there is a strong interconnection between topic discovery and emotion detection. Jointly modeling topics and emotions is a suitable strategy for these tasks. This book also examines the impact of topics on emotion detection and other related areas.

As a well-known text mining tool, topic modeling can effectively discover the latent semantic structure of text data. Extracting topics from documents is also one of the fundamental challenges in natural language processing. Although topic models have seen significant achievements over the past three decades, there remains a scarcity of methods that effectively model temporal aspect. Moreover, many contemporary topic models continue to grapple with the issue of noise contamination, particularly in social media data.This book presents several approaches designed to address these two limitations. Initially, traditional lifelong topic models aim to accumulate knowledge learned from experience for future task. However, the sequence of topics extracted by these methods may shift over time, leading to semantic misalignment between the topic representations across document streams. Such misalignment can degrade the performances of various downstream tasks, including online document classification and dynamic information retrieval at the topic level. Additionally, the challenge of coherent topic modeling is particularly relevant due to the noise and large scale of social media datasets. Messages on social media platforms often consists of only a few words, resulting in a lack of significant context. Models applied directly to this type of text frequently encounter the problem of feature sparsity, which can yield unsatisfactory outcomes.In the context of emotion detection, public emotions are known to fluctuate across different topics, and topics can evoke public emotion. Thus, there is a strong interconnection between topic discovery and emotion detection. Jointly modeling topics and emotions is a suitable strategy for these tasks. This book also examines the impact of topics on emotion detection and other related areas.

This textbook offers a comprehensive and accessible introduction to machine learning with the Julia programming language. It bridges mathematical theory and real-world practice, guiding readers through both foundational concepts and advanced algorithms. Covering topics from essential principles like Kullback–Leibler divergence and eigen-analysis to cutting-edge techniques such as deep transfer learning and differential privacy, each chapter delivers clear explanations and detailed algorithmic treatments. Sample code accompanies every major topic, enabling hands-on learning and faster implementation.By leveraging Julia’s powerful machine learning ecosystem—including libraries such as Flux.jl, MLJ.jl, and more—this book empowers readers to build robust, state-of-the-art machine learning models.Ideal for students, researchers, and professionals alike, this textbook is designed for those seeking a solid theoretical foundation in machine learning, along with deep algorithmic insight and practical problem-solving inspiration.

Today, machine learning algorithms are often distributed across multiple machines to leverage more computing power and more data. However, the use of a distributed framework entails a variety of security threats. In particular, some of the machines may misbehave and jeopardize the learning procedure. This could, for example, result from hardware and software bugs, data poisoning or a malicious player controlling a subset of the machines. This book explains in simple terms what it means for a distributed machine learning scheme to be robust to these threats, and how to build provably robust machine learning algorithms.Studying the robustness of machine learning algorithms is a necessity given the ubiquity of these algorithms in both the private and public sectors. Accordingly, over the past few years, we have witnessed a rapid growth in the number of articles published on the robustness of distributed machine learning algorithms. We believe it is time to provide a clear foundation to this emerging and dynamic field. By gathering the existing knowledge and democratizing the concept of robustness, the book provides the basis for a new generation of reliable and safe machine learning schemes. In addition to introducing the problem of robustness in modern machine learning algorithms, the book will equip readers with essential skills for designing distributed learning algorithms with enhanced robustness. Moreover, the book provides a foundation for future research in this area.

Today, machine learning algorithms are often distributed across multiple machines to leverage more computing power and more data. However, the use of a distributed framework entails a variety of security threats. In particular, some of the machines may misbehave and jeopardize the learning procedure. This could, for example, result from hardware and software bugs, data poisoning or a malicious player controlling a subset of the machines. This book explains in simple terms what it means for a distributed machine learning scheme to be robust to these threats, and how to build provably robust machine learning algorithms.Studying the robustness of machine learning algorithms is a necessity given the ubiquity of these algorithms in both the private and public sectors. Accordingly, over the past few years, we have witnessed a rapid growth in the number of articles published on the robustness of distributed machine learning algorithms. We believe it is time to provide a clear foundation to this emerging and dynamic field. By gathering the existing knowledge and democratizing the concept of robustness, the book provides the basis for a new generation of reliable and safe machine learning schemes. In addition to introducing the problem of robustness in modern machine learning algorithms, the book will equip readers with essential skills for designing distributed learning algorithms with enhanced robustness. Moreover, the book provides a foundation for future research in this area.

Unsupervised domain adaptation (UDA) is a challenging problem in machine learning where the model is trained on a source domain with labeled data and tested on a target domain with unlabeled data. In recent years, UDA has received significant attention from the research community due to its applicability in various real-world scenarios. This book provides a comprehensive review of state-of-the-art UDA methods and explores new variants of UDA that have the potential to advance the field.The book begins with a clear introduction to the UDA problem and is mainly organized into four technical sections, each focused on a specific piece of UDA research. The first section covers criterion optimization-based UDA, which aims to learn domain-invariant representations by minimizing the discrepancy between source and target domains. The second section discusses bi-classifier adversarial learning-based UDA, which creatively leverages adversarial learning by conducting a minimax game between the feature extractor and two task classifiers. The third section introduces source-free UDA, a novel UDA setting that does not require any raw data from the source domain. The fourth section presents active learning for UDA, which combines domain adaptation and active learning to reduce the amount of labeled data needed for adaptation.This book is suitable for researchers, graduate students, and practitioners who are interested in UDA and its applications in various fields, primarily in computer vision. The chapters are authored by leading experts in the field and provide a comprehensive and in-depth analysis of the current UDA methods and new directions for future research. With its broad coverage and cutting-edge research, this book is a valuable resource for anyone looking to advance their knowledge of UDA.

Unsupervised domain adaptation (UDA) is a challenging problem in machine learning where the model is trained on a source domain with labeled data and tested on a target domain with unlabeled data. In recent years, UDA has received significant attention from the research community due to its applicability in various real-world scenarios. This book provides a comprehensive review of state-of-the-art UDA methods and explores new variants of UDA that have the potential to advance the field.The book begins with a clear introduction to the UDA problem and is mainly organized into four technical sections, each focused on a specific piece of UDA research. The first section covers criterion optimization-based UDA, which aims to learn domain-invariant representations by minimizing the discrepancy between source and target domains. The second section discusses bi-classifier adversarial learning-based UDA, which creatively leverages adversarial learning by conducting a minimax game between the feature extractor and two task classifiers. The third section introduces source-free UDA, a novel UDA setting that does not require any raw data from the source domain. The fourth section presents active learning for UDA, which combines domain adaptation and active learning to reduce the amount of labeled data needed for adaptation.This book is suitable for researchers, graduate students, and practitioners who are interested in UDA and its applications in various fields, primarily in computer vision. The chapters are authored by leading experts in the field and provide a comprehensive and in-depth analysis of the current UDA methods and new directions for future research. With its broad coverage and cutting-edge research, this book is a valuable resource for anyone looking to advance their knowledge of UDA.