Deep Neural Networks in a Mathematical Framework

This SpringerBrief provides a rigorous mathematical framework for deep neural networks, derived gradient descent algorithms, and more concise and intuitive representation. It aims to unlock the black box of Deep Learning and catalyze further discoveries.

Format: Paperback / softback
Length: 84 pages
Publication date: 03 April 2018
Publisher: Springer International Publishing AG

This SpringerBrief provides a comprehensive guide on how to develop a rigorous end-to-end mathematical framework for deep neural networks. The authors offer valuable tools to represent and describe neural networks, shedding light on previous findings in the field in a more natural and intuitive manner. By deriving gradient descent algorithms in a unified way for various neural network structures, including multilayer perceptrons, convolutional neural networks, deep autoencoders, and recurrent neural networks, the authors offer a more concise and mathematically intuitive representation of neural networks. This SpringerBrief is a significant step towards unraveling the mysteries of deep learning, with the authors believing that this framework will foster further advancements in understanding the mathematical properties of neural networks. Accessible to researchers, professionals, and students in the field of deep learning as well as those outside of the neural network community, this SpringerBrief is a valuable resource for anyone seeking to deepen their understanding of this rapidly evolving field.

Weight: 170g
Dimension: 156 x 234 x 14 (mm)
ISBN-13: 9783319753034
Edition number: 1st ed. 2018