Deep learning

Deep learning includes architectures and techniques for supervised and unsupervised learning with the capacity to internalize the abstract structure of high-dimensional data using networks composed of building blocks to create discriminative or generative models. These techniques have proved enormously successful in recent years and any reader interested in mastering them must become familiar with the basic building blocks of deep learning first and understand the various types of networks in use by practitioners. Hands-on experience building and tuning deep neural networks is invaluable if you intend to get a deeper understanding of the subject. Deep learning, in various domains such as image classification and text learning, incorporates feature generation in its structures thus making the task of mining the features redundant in many applications. The following sections provide a guide to the concepts, building blocks, techniques for composing architectures, and training...