Deep learning is a popular topic these days since it is a high-level machine learning approach that combines a class of learning algorithms with the application of multiple layers of nonlinear units. A subfield of machine learning, deep learning focuses on powerful algorithms inspired by the structure and function of the human brain called neural networks. A unified geometric analysis to understand the working mechanism of deep learning from high-dimensional geometry is offered. Then, different forms of generative models like GAN, VAE, normalizing flows, optimal transport, and so on are described from a unified geometric perspective, showing that they actually come from statistical distance-minimization problems. This textbook provides a clear and interesting introduction to deep learning, covering a variety of connectionist models that are considered state-of-the-art at the moment. Using a clear and comprehensible writing style, the work examines today's most well-liked algorithms and architectures while also providing a thorough breakdown of their underlying mathematical foundations. Convolutional networks, long short-term memories, word2vec, random bit networks, decision-based networks, neural turing machines, memory networks, autoencoders, and more are all discussed.
