Jawaharlal Nehru Technological University Anantapur 2010-1st Sem B.Tech Code :R7411009 3 IV (R07) Regular s, ember/ember ARTIFICIAL NEURAL NETWORKS - Question Paper

Code :R7411009 3
IV B.Tech I semester (R07) Regular Examinations, November/December 2010
ARTIFICIAL NEURAL NETWORKS
(Electronics & Instrumentation Engineering)
Time: three hours Max Marks: 80
ans any 5 ques.
All ques. carry equal marks
1. (a) De¯ne activation functions, bias, threshold and learning in situation to Arti¯cial Neural
Networks.
(b) discuss models of arti¯cial neural networks feed forward and feedback networks.
2. (a) discuss outstar learning rule.
(b) In a neural nerwork, a neuron is fed with 4 input models. The 3 input sets are
X1 = £ one ¡2 0 one ¤T
;X2 = £ 0 1:5 ¡0:5 ¡1 ¤T
;
X3 = £ ¡1 one 0:5 ¡1 ¤T
:
Weights are initialized to W1 = £ one ¡1 0 0:5 ¤T
: The learning rate constant c=0.1.
The teacher's desired responses for x1,x2,x3 are d1=-1 ,d2=-1 ,d3=1 respectively.
Determine the weights after training the neural network using perceptron rule for one
cycle.
3. (a) discuss fault back propagation algorithm.
(b) Di®erentiate ranging from back propagation and radical basis function network.
4. (a) explain in detail the MRII training algorithm.
(b) With architecture, discuss the MRI training algorithm.
5. (a) discuss outstar learning.
(b) discuss Learning Vector Quantizer Architecture and its algorithm.
6. (a) discuss architecture of continuous Hop¯eld network.
(b) discuss the storage and retrival algorithms of Hop¯eld network.
7. (a) discuss Boltzman machine architecture.
(b) What is the basic concept behind ART?
8. discuss in detail the implementation of A/D converter using Hop¯eld network.

Earning:   Approval pending.