Scala for Machine Learning - Second Edition

Chapter 11. Deep Learning

This chapter leverages the concepts and components of the multilayer perceptron described in the previous chapter and applies them to deep learning architectures. There has been quite a bit of buzz surrounding deep learning lately, although the algorithms presented in this chapter were introduced 20 to 30 years ago.

The recent advance in neural networks has as much to do with the availability of powerful hardware such as memory-based distributed computing and GPU as the academic research.

This chapter describes the following:

• Sparse autoencoders as a dimension reduction technique for non-linear problems
• Binary restricted Boltzmann machines as the core foundation of deep generative models for unsupervised learning
• Convolutional ...