Go up to Question 2
Solution to Question 2
- h1 represents the Boolean function "short and (known or
new)" (which is the same function as the decision tree of Figure
11.3).
- h2 represents the Boolean function "not known and not new
and not short".
- reads represents the Boolean function "h1". That is, reads is
true if h1 is true and is false otherwise.
Thus the neural network makes the same predictions on both seen and
unseen examples as the decision tree!