Название: Deep Learning with R Автор: Abhijit Ghatak
Издательство: Springer
ISBN: 9811358494
Год: 2019
Страниц: 259
Язык: английский
Формат: pdf (true), epub
Размер: 59.9 MB
Deep Learning with R introduces deep learning and neural networks using the R programming language. The book builds on the understanding of the theoretical and mathematical constructs and enables the reader to create applications on computer vision, natural language processing and transfer learning.
Deep learning is a specific subfield of machine learning, which makes use of tens/hundreds of successive layers of representations. The specification of what a layer does to its input is stored in the layer’s parameters. Learning in deep learning can also be defined as finding a set of values for the parameters of each layer of a deep learning model, which will result in the appropriate mapping of the inputs to the associated answers (outputs). Deep learning has been proven to be better than conventional machine learning algorithms for these “perceptual” tasks, but not yet proven to be better in other domains as well.
The book starts with an introduction to machine learning and moves on to describe the basic architecture, different activation functions, forward propagation, cross-entropy loss and backward propagation of a simple neural network. It goes on to create different code segments to construct deep neural networks. It discusses in detail the initialization of network parameters, optimization techniques, and some of the common issues surrounding neural networks such as dealing with NaNs and the vanishing/exploding gradient problem. Advanced variants of multilayered perceptrons namely, convolutional neural networks and sequence models are explained, followed by application to different use cases. The book makes extensive use of the Keras and TensorFlow frameworks.
This book requires an intermediate level of skill in R and no previous experience of deep learning.
Contents:
Скачать Deep Learning with R