Punjab Technical University 2005 M.Tech Electronics and Communication Engineering NEURAL NETWORK AND FUZZY LOGIC ECE 505 - Question Paper

NEURAL NETWORK AND FUZZY LOGIC ECE 505 first Sem May 2k5

Max marks 100

Note: Attempt any 5 ques.. All ques. carry equal marks

1. Implement the backpropagation learning algorithm for a fully connected 3 layer network. Include parameter for layer sizes, learning rate, number of training epochs.

2. (a) What do you understand by learning? elaborate the major characteristics of every of them?
(b) discuss the Kohonen’s self-organising network

3. What are the features of counter propagation network? How are these features achieved?

4. How neural network approach can be applied to VLSI design? elaborate the algorithms which can be used in that? discuss.

5. (a) What do you understand by perception? discuss the fixed increment perception learning algorithm.

(b) discuss the random adaptive bidirectional associative memories. State the RABAM theorem.

6. (a) elaborate the linguistics variables.
(b0 elaborate the different kinds of membership functions? discuss with examples.

7. (a) elaborate the different steps in the design of a fuzzy system? discuss with examples.

(b) How PID control can be designed using fuzzy logic?

8. Write short notes on the following:
(a) K mean clustering algorithm
(b) History of neural network

Earning:   Approval pending.