
	Trending ▼ ICSE CBSE 10th ISC CBSE 12th CTET GATE UGC NET Vestibulares ResFinder

2003 Course Artificial Neural Networks

4 pages, 39 questions, 0 questions with responses, 0 total responses,

	pune_eng	+Fave Message
	Profile Timeline Uploads Q & A

Home > pune_eng >

Instantly get Model Answers to questions on this ResPaper. Try now!
NEW ResPaper Exclusive!

Formatting page ...

Total No. of Questions :12] [Total No. of Pages :4 P1076 [3864]-268 B.E. (E&T/C) ARTIFICIAL NEURAL NETWORKS (ANN) (404218) (2003 Course) Time : 3 Hours] [Max. Marks : 100 Instructions to the candidates: 1) 2) 3) 4) 5) 6) Answer Q. 1 or Q. 2, Q. 3 or Q. 4, Q. 5 or Q. 6 from section - I, Q. 7 or Q. 8, Q.9 or Q. 10, Q. 11 or Q. 12 from section - II. Answers to the two sections should be written in separate books. Neat diagrams must be drawn wherever necessary. Figures to the right indicate full marks. Use of electronic pocket calculator is allowed. Assume suitable data, if necessary. SECTION - I Q1) a) b) c) Explain biological neural network with neat diagram and compare its performance with artificial neural network. [8] A neural network has 2 - input neurons with the following weight matrix W = [3 2] and I = [-5 7]t and the required output is 0.5. i) Find out the bias that will do the job, if linear transfer function is used. ii) Find out the bias, that will do the job, if log sigmoid function is used. [4] Draw and explain the following structures of ANN. [6] i) Group of instars. ii) Group of out stars. iii) Auto associate memory. OR Q2) a) Design OR gate using MP neuron model. Explain the drawbacks of MP neuron model. [4] b) Explain the difference between activation dynamic models and synaptic dynamic models. [6] P.T.O. c) Explain the difference between stability and convergence. [4] d) What is supervised learning? Mention the different supervised algorithms. [4] Q3) a) If I1= [ 1 1 ] t, t1 = 1 and I2 = [ 1 -1 ] t, t2 = -1 represents the input and output target pairs. find out the new weights using perceptron learning for one epoch. Initial weight = [ 0 0], learning rate = 0.5 [6] b) Explain the term linear separability. Explain wheter ExoR gate is linearly separable or not. [4] c) Explain the training algorithm used in adaline and explain how the decision boundary of adaline differs from that of perceptron learning.[6] OR Q4) a) Explain the training algorithm used in BPN (Back propogation network). Explain how BPN with momentum differs from that of BPN. [8] b) A back propogation network with 2i/p neurons, 2 hidden neurons and 1 output neuron is given. Input X = [ 0.1 0.7 ] t = 1 learning rate = 1. Weight matrix between input and hidden layer is W. 0.1 0.3 W= 0.2 0.5 Q5) a) Weight matrix between hidden and output layer is V V = [ 0.2 0.1 ] Find the new weight matrix W and V after 1 epoch of training. [8] Design a Hopfield network for 5 bit bipolar patterns. The training patterns are S1 = [ 1 1 1 1 1 ], S2 = [ 1, -1, -1, 1, -1 ] S3 = [ -1 1 -1 -1 -1 ]. Find the weight matrix and energy for three samples. [8] b) c) Explain what is feedback network. [4] Explain the concept of simulated annealing and explain how it is used in neural networks. [4] [3864]-268 -2- OR Q6) a) Draw the architecture of Boltzmann machine and explain how it is different from Hopfield network. [6] b) Explain the term energy associated with Hopfield n/w. [2] c) What is a state transition diagram for a feedback network? Explain how to derive it for a given network. [8] SECTION - II Q7) a) Explain the architecture of self organised feature map (SOFM) network. Explain the training algorithm used for the same. [8] b) Explain the architecture and application of Maxnet in unsupervised learning. [6] c) What is learning vector quantization? [2] OR Q8) a) Explain ART1 architecture. [4] b) Explain the training algorithm used in ART1 network. [8] c) Explain plasticity stability dilemma. [4] Explain TAM and MAM with the help of neat diagrams. [6] Q9) a) b) Explain the principle of Neocognitron for pattern variability tasks. c) Explain the architecture of RBF network with the help of neat diagram. [4] OR [3864]-268 -3- [6] Q10) a) Design a BAM for the following input and output pairs. i) Find the weight matrix. ii) [10] Verify the operation of BAM in X and Y direction. I1 = ( 1, -1, -1, 1 ) t1 = (-1, 1, 1) I2 = (-1, 1, 1, -1 ) t2 = (1, -1, 1) I1and I2 are inputs. t1and t2 are target outputs. b) Explain the architecture of hetero associative net and explain the training algorithm used for the same. [6] Q11) a) What is the problem in the recognition of hand written digits? [6] b) Explain the difficulties in the solution of travelling salesman problem by a feedback neural network. [8] c) Explain how an image smoothing problem can be solved by principles of neural network. [4] OR Q12) a) Explain how neural network principles are useful for a texture classification problem. [6] b) Explain how neural network principles are useful in control applications. [6] c) What is the significance of neural networks in the NET talk application. [6] nnn [3864]-268 -4-

Formatting page ...

Tags : Pune, Engineering, University of Pune, Engineering question papers, Pune University, previous year question papers, question papers, india, model question paper, pune university paper pattern, pune university syllabus, old question papers

pune_eng chat

Horizontal lines at:
Guest Horizontal lines at:
AutoRM Data:
Box geometries: