Punjab Technical University 2001 B.Tech Production Engineering EC-440 NEURAL NETWORK

EC-440 NEURAL NETWORK & FUZZY LOGIC /2K1}

(B.Tech., eighth Semester, 2125)

Time: three Hours

Maximum Marks: 60

Note: part A is compulsory. Attempt any 4 ques. from part B
and any 2 ques. from part C.


1. part - A

(a) What is Recurrent Networks architecture of Neural Networks?

(b) discuss competitive learning.

(c) What is step function, signum function?

(d) discuss different kind of non-linear activation operators?

(e) discuss Basic fuzzy set operations?

(f) Write about fuzzy Quantifier classes?

(g) explain Neural Network classification.

(h) Which models are called casual models?

(i) Write objectives to train supervised learning?

(j) What is threshold purning?
(Marks 10x2=20)

part - B

2.What is modus ponens & chain Rules in propositional logic? discuss with suitable example.

3.What is training time, predictive power of supervised learning? discuss convergence in probability?

4.Compare fault correction learning & memory based learning?

5.Explain counter propagation networking & compare with ART network.

6.Compare hebb's Rule Delta learning rules.

( Marks four x five =20)
part - C

7.Write all the 6 basic learning tasks that a neural network needed to perform. discuss in detail.

8.Explain atleast 4 various forms of recurrent networ~n detail & how they are co-related.

9.Write & compare centroid method & mean of maxima of defuzzification in fuzzy loge?

(Marks 2x10=20)

Earning:   Approval pending.