Artificial Neural Networks

This talk is based on

• Tom M. Mitchell. Machine Learning. McGraw Hill. 1997. Chapter 4.
• and his slides.

1. 1 Introduction
   1. 1.1 The Human Brain
   2. 1.2 Neural Network Representation
2. 2 When to Use Neural Networks
   1. 2.1 ALVINN
3. 3 Perceptrons
   1. 3.1 Representational Power of Perceptrons*
   2. 3.2 Perceptron Training
   3. 3.3 Perceptron Training Rule Convergence
   4. 3.4 Gradient Descent
      1. 3.4.1 Gradient Descent Landscape
      2. 3.4.2 Calculating the Gradient Descent
      3. 3.4.3 Gradient Descent Algorithm
   5. 3.5 Perceptron Learning Summary, so far
   6. 3.6 Incremental (Stochastic) Gradient Descent
      1. 3.6.1 Stochastic versus Batch Gradient Descent
4. 4 Multilayer Networks
   1. 4.1 Sigmoid Unit
   2. 4.2 Error Gradient for Sigmoid Unit
   3. 4.3 Backpropagation
      1. 4.3.1 Backpropagation Details
      2. 4.3.2 Hidden Layer Representation*
      3. 4.3.3 8-3-8 Plots
      4. 4.3.4 Backpropagation Convergence
   4. 4.4 Representational Power of ANNs
   5. 4.5 Overfitting ANNs
5. 5 Face Recognition Example
6. 6 Alternative Error Functions
7. 7 Recurrent Networks
8. 8 Dynamically Modifying Network Structure
9. 9 Summary

Entire Presentation with Notes

---

Copyright © 2009 José M. Vidal . All rights reserved.

01 February 2003, 09:41AM