Homework 4 - Chapters 6 and 7

Each question is worth 2 points.

1. What is a recurrent network? Describe how its learning differs from a feedforward network.
2. How do Hopfield networks and bidirectional associative memory differ?
3. Show all steps of recalling the associated pattern for X = [ -1 1 -1] in the Hopfield network presented on pages 190-192.
4. How do self-organized maps differ from feedforward and recurrent neural networks?
5. What are the differences between Hebbian learning and competitive learning?
6. Why is the forgetting factor included in the generalized activity product rule for Hebbian learning?
7. 7.1: Why are genetic algorithms called genetic? Who is the 'father' of genetic algorithms?
8. Describe the main steps of a genetic algorithm.
9. Show the result of a single-point crossover in the middle of the bit strings 0011 1100 and 1111 0000.
10. What is the purpose of the mutation operator? Is it always beneficial? Explain why or why not.