Bindning: Häftat band
Bokförlag: Borås Studentbokhandel
ISBN: 9789198346657
Omfång: 292 s.
Språk: Engelska
Utgivningsdatum: 2025-09-15
Förlagets information
About the Author Magnus Bengtsson is an Associate Professor (Docent) at the University of Borås, Sweden, where he conducts research and teaching in machine learning. With more than 20 years of experience in applied projects across industry and academia, he has distilled his insights into this course book, designed to explain machine learning models from the ground up to the state-of-the-art architectures used today. About the Book Machine Learning – From Foundations to Advanced Architectures The field of machine learning can feel overwhelming, but it is easier than you might think. What you need is a “Rosetta Stone” to connect mathematics, theory, and code into clear understanding and working systems. This book provides such a guide. It begins with fundamental concepts in calculus, linear algebra, numerical methods, and optimization, then gradually builds toward neural networks, convolutional architectures, transformers, GANs, spiking neural networks, and other state-of-the-art models. Balancing conceptual clarity with practical coding, the book takes a hands-on, from-scratch approach: setting up the environment, defining datasets, configuring training and validation, and avoiding common version pitfalls. Each chapter combines explanations, derivations, and runnable Python code. Designed for students, engineers, and researchers, this book is both a foundation and a roadmap—helping you understand the principles of machine learning while equipping you to implement models that work in practice.
Machine Learning – From Foundations to Advanced Architectures The field of Machine Learning can be overwhelming, and you might feel like you're losing control. Instead of giving up, let me share a secret with you: it's easier than you think. All you need is a kind of “Rosetta Stone” to fully accelerate your work with machine learning. This book offers such a guide. It takes you on an alternative route, starting with fundamental concepts from calculus, linear algebra, numerical methods, and optimization, and leading up to the state-of-the-art algorithms that have emerged over the last couple of decades. My background in industry, research, and teaching has given me deep insights into the common challenges of developing efficient algorithms for prediction. That experience is distilled here into a coding-oriented, “from-scratch” approach, where we focus on setting up the environment properly, defining datasets, configuring training and validation, and avoiding version-related pitfalls that often cause problems in real-world projects. Balancing conceptual clarity with practical implementation, the book combines detailed derivations, clear illustrations, and runnable Python code. It is designed for students, engineers, and researchers who want to not only understand the principles of machine learning but also build systems that work. Contents (main chapters): • Nomenclature • Foreword • Introduction • Numerical Methods – The Basis for ML • Neural Networks (ANN) • ResNet • Autograd • Convolutional Neural Networks (CNN) • Activation Functions • System Analysis – LeNet-5 Case Study • Transformers • Discrete Wavelet Transform • Support Vector Machines (SVM) • Principal Component Analysis (PCA) • Generative Adversarial Networks (GAN) • Spiking Neural Networks (SNN) • Equivariant CNNs • Data Structures for Object Detection • Glossary of AI and Machine Learning Terms • Bibliography
