دانلود کتاب دوره فشرده یادگیری عمیق -مقدمه‌ای عملی و پروژه‌محور بر هوش مصنوعی

The deep learning revolution has arrived, and it’s changing our lives in ways we never imagined. From everyday conveniences to groundbreaking innovations, deep learning is at the core of many of the technologies shaping our future. Whether you’re ordering food, planning your commute, or searching for a playlist, deep learning is optimizing these experiences behind the scenes. But its impact goes far beyond these everyday tasks. Today, deep learning is enabling scientists to design drugs tailored to specific receptors, accelerating medical breakthroughs and transforming health care. Large language models (LLMs) like GPT-4 are capable of generating text and writing code, revolutionizing fields like education, programming, and research. Deep learning is even reshaping social interactions, sometimes in unsettling ways, such as AI-powered chatbots serving as virtual companions or romantic partners. Whether you need to learn deep learning for your career or are just curious about how it works, this book is for you. It covers everything from basic concepts to advanced problem-solving tools, equipping you to understand deep learning so that you can go from being a passive user to someone who can control and optimize this technique for your own unique needs. Although the terms deep learning, machine learning, and artificial intelligence are often used interchangeably, they’re different. Artificial intelligence (AI) refers to any technique used to create systems capable of performing cognitive tasks that usually require human intelligence, such as recognizing faces, understanding speech, making complex decisions, or translating languages. Examples of AI that don’t involve machine learning include rule-based systems and expert systems used in medical diagnosis (remember IBM Watson?). Machine learning is a branch of AI where computers use algorithms to learn from data and make predictions. Importantly, not all forms of machine learning fall under deep learning—for example, the decision trees and support vector machines that were popular in the early 2000s. In turn, deep learning is a branch of machine learning that uses neural networks, typically with many layers of artificial neurons (hence the term deep). The ability of deep learning to learn directly from data makes it effective for solving a wide range of problems, from recognizing objects in images and videos, to translating spoken and written languages, to playing complex strategy games like Go. Notable applications encompass advanced conversational agents like ChatGPT, image-generation models like DALL-E, and protein structure prediction tools like AlphaFold. With its foundation in neural network research from the 1950s to the 1980s, deep learning gained momentum after breakthroughs in the early 2000s and exploded in popularity in the mid-2010s. This revolution was fueled by gamers and social networks. As increasingly powerful graphics processing units (GPUs) were developed for gaming, social media provided a gold mine of data for training models. The turning point came in 2012 when a deep learning model, AlexNet, won the ImageNet Large Scale Visual Recognition Challenge. ImageNet is a large database of annotated images. The competition involves developing models to accurately classify and detect objects within these images. AlexNet’s success demonstrated the potential of deep learning, leading to its widespread adoption in just a few years. The rapid development of AI brought about by these transformations has led to all sorts of sensational conjectures: the utopian remaking of society, vast economic disruption, and even humanity’s extermination. In this environment, understanding AI has never been more crucial, and the best way to grasp its nuances is through direct experience. This direct experience should also convince you that, despite predictions of doom and gloom, the most creative and purposeful aspects of our lives and careers will likely remain uniquely human for a long time. This book is designed for those eager to dive into deep learning through practical, hands-on experience, without assuming any prior knowledge of deep learning. Each chapter is structured around examples and projects that are computationally manageable on standard hardware yet realistic in scope. Even so, readers will understand and implement the latest advancements in the field, such as diffusion models for generating high-resolution images and vision transformers (ViTs) for image classification.