Lecture Notes In Deep Learning eBook

The compendium provides an introduction to the theory of deep learning, from basic principles of neural network modeling and optimization to more advanced topics of neural networks as Gaussian processes, neural tangent and information theory.

This unique reference text complements a largely missing theoretical introduction to neural networks without being overwhelmingly technical in a level accessible to upper-level undergraduate engineering students.

Advanced chapters were designed to offer an additional intuition into the field by explaining deep learning from statistical and information theory perspectives. The book further provides additional intuition to the field by relating it to other statistical and information modeling approaches.

Contents:

• Preface
• Neural Network Basics:
  • Introduction
  • Neural Networks in Use
  • Optimization Methods
• Representation Learning:
  • Autoencoders and Principal Components Analysis
  • Probabilistic Principal Components Analysis and Variational Autoencoder
  • Variational Autoencoder
• Convolutional, Recurrent, and Transformer Neural Networks:
  • Convolutional Neural Networks
  • CNN Applications in Vision and Audio
  • Recurrent Neural Network
  • Attention and Transformers
• Generative Models:
  • Generative Adversarial Networks
  • Wasserstein Generative Adversarial Network
  • Normalizing Flows and Diffusion Models
• Deeper Understanding of Deep Learning:
  • Information Theory of Learning
  • NN as Gaussian Processes
• Further Topics:
  • Transfer Learning
  • Explainable AI
  • Deep Reinforcement Learning
• Conclusion
• Bibliography
• Index

Readership: Researchers, professionals, academics, and undergraduate and graduate students in artificial intelligence and data bases/info science.