Meeting 23: Learning Theory

Reading: AIAMA 18.4-18.5

There are some basic principles that are common to all learning methods. Now that we have seen an example of supervised learning we have some context to help explain this theory.

Questions you should be able to answer after reading are:

  1. What are the assumptions about the target function (model) to be learned and the data available for training and testing?
  2. What is the difference between the training and test set?
  3. What is cross-validation?
  4. What does it mean to select the best model, what are the criteria used?
  5. what does probably approximately correct mean?