Exercises

1. How appropriate would the Rote learning methods of Samuel's Checkers program be to Chess?
2. Write semantic nets descriptions of the Blocks World tent and arch structures in Winston's Learning Program. Figure out some near miss structures for all structures and describe them as semantic nets.
3. How may two arch structure be compared in Winston's Learning Program?
4. Write a decision tree description of the flush in poker example.
5. Compare how version spaces and decision trees learn with the methods of neural networks and genetic algorithms. How sensitive is each method to the training data supplied?
6. Show how version spaces could be used to learn the concept of pairs (two of a kind) in a game such as five card poker. (You can supply your own learning set but make sure you include appropriate positive and negative example).
7. Discover how AM discovered prime numbers.