دانلود کتاب یادگیری عمیق

Deep learning uses multilayered neural networks trained with large data sets to solve complex information processing tasks and has emerged as the most successful paradigm in the field of machine learning. Over the last decade, deep learning has revolutionized many domains including computer vision, speech recognition, and natural language processing, and it is being used in a growing multitude of applications across healthcare, manufacturing, commerce, finance, scientific discovery, and many other sectors. Recently, massive neural networks, known as large language models and comprising of the order of a trillion learnable parameters, have been found to exhibit the first indications of general artificial intelligence and are now driving one of the biggest disruptions in the history of technology. This expanding impact has been accompanied by an explosion in the number and breadth of research publications in machine learning, and the pace of innovation continues to accelerate. For newcomers to the field, the challenge of getting to grips with the key ideas, let alone catching up to the research frontier, can seem daunting. Against this backdrop, Deep Learning: Foundations and Concepts aims to provide newcomers to machine learning, as well as those already experienced in the field, with a thorough understanding of both the foundational ideas that underpin deep learning as well as the key concepts of modern deep learning architectures and techniques. This material will equip the reader with a strong basis for future specialization. Due to the breadth and pace of change in the field, we have deliberately avoided trying to create a comprehensive survey of the latest research. Instead, much of the value of the book derives from a distillation of key ideas, and although the field itself can be expected to continue its rapid advance, these foundations and concepts are likely to stand the test of time. For example, large language models have been evolving very rapidly at the time of writing, yet the underlying transformer architecture and attention mechanism have remained largely unchanged for the last five years, while many core principles of machine learning have been known for decades.