Nov 3
Skim through and discuss the Naive Bayes example in unit 5 from ai-class.com, including their laplace smoothing and over-fitting prevention scheme.
Discuss linear regression and fitting generally and what it has to do with machine learning; this is a start towards neural networks.
- "loss" function
- iterative "direct descent" approach
- how does this relate to the search methods we did before?
Start discussing the "perceptron"; a simplistic linear neural net.
- weights
- training data
- "back" propagation of corrections
