Question Bank

ANN

Unit I
Short Question
1. Define ANN and neural computing?
Ans - Neural networks, also known as artificial neural networks (ANNs) or simulated neural networks
(SNNs), are a subset of machine learning and are at the heart of deep learning algorithms. Their name and
structure are inspired by the human brain, mimicking the way that biological neurons signal to one another.

2. Mention the characteristics of problems suitable for ANNs?

Ans - The target function output may be discrete-valued, real-valued, or a vector of several real-valued or
discrete-valued attributes. The training examples may contain errors. Long training times are acceptable. Fast
evaluation of the learned target function may be required.

3. List some applications of ANNs?

Ans - Artificial neural networks are used for a range of applications, including image recognition, speech
recognition, machine translation, and medical diagnosis. The fact that ANN learns from sample data sets is a
significant advantage. The most typical application of ANN is for random function approximation.

4. What are neurons?

Ans - Neurons (also called neurones or nerve cells) are the fundamental units of the brain and nervous
system, the cells responsible for receiving sensory input from the external world, for sending motor
commands to our muscles, and for transforming and relaying the electrical signals at every step in between.

Long Question
1. Explain the biological prototype of neuron. Also explain the characteristics of neuron.
2. Describe an artificial neuron and discuss its functioning. Compare it with biological neuron?
3. Explain the Cerebrum and Cerebellum in detail?
4. What are different Light Sensing organs, explain in detail?
5. List the application of ANN and explain any two in details?
6. Explain the biological neural network?
7. Explain nervous system in details?
8. Different the CNS and PNS?
9. How the information is processed in the nervous system?
10. What are the different components of the neurons, explain in details?

Unit II

Short Question
1. Specify the learning law used by each ANN.
Ans - Hebbian learning rule – It identifies, how to modify the weights of nodes of a network. Perceptron
learning rule – Network starts its learning by assigning a random value to each weight. Delta learning
rule – Modification in sympatric weight of a node is equal to the multiplication of error and the input.

2. What is ANN? List the characteristics of ANN

Ans - A neural network is a method in artificial intelligence that teaches computers to process data in a
way that is inspired by the human brain. It is a type of machine learning process, called deep learning,
that uses interconnected nodes or neurons in a layered structure that resembles the human brain.

3. Explain in brief basic models of Artificial Neural Network.

 Ans -

4. Define Bias?
Ans- It is a constant value that allows to shift activation function to get desired output.It helps the NN
model which helps can best fit for input pattern.

5. Define Weight?
Ans-Weight is the parameter within a neural network that transforms input data within the network's
hidden layers. A neural network is a series of nodes, or neurons.

Long Question
1. Discuss the methods, which have been developed to improve generalization of neural network learning.
2. Activation function involved in the computation backpropagation
3. Rate of learning in back propagation algorithm.
4. Write history of artificial neural system development.
5. List and explain the various activation functions used in ANN.
6. Discuss the following major components of a generalized ANN model: processing units; states of
activation; output signals of the units; patterns of connectivity; and activation rule?
7. Draw and explain neural network based OR function?
8. Calculate the output of the neuron Y for the network shown in the figure using the activation function
as (a) binary sigmoid (b) bipolar sigmoid (c) Identity activation
 9. Define and explain delta learning rule?
10. Design a Hebbian neuron/net to implement the logic 2 input AND function using bipolar input-output
pattern?

Unit III

Short Question
1. Give some examples for Nonrecurrent and Recurrent ANNs?
Ans - Google, Bing, and DuckDuckGo are the most prominent examples of recurrent neural network image
recognition. The goal is to images that fit the input query or find images that look like an input image.

2. What is the difference between Forward propagation and Backward Propagation in Neural Networks?
Ans - Backward Propagation is the process of moving from right (output layer) to left (input layer). Forward
propagation is the way data moves from left (input layer) to right (output layer) in the neural network. A
neural network can be understood by a collection of connected input/output nodes.

3. Differentiate between Jordan and Elman Networks?

Ans - The only difference between the Jordan and Elman RNNs lies in the factors c|O| and c|H|. Their
difference in complexity depends thus on the size of the output layer (Jordan) with respect to the size of the
hidden layer (Elman).

4. What do you mean by selecting learning rate?

Ans - Learning rate is used to scale the magnitude of parameter updates during gradient descent. The choice
of the value for learning rate can impact two things: 1) how fast the algorithm learns and 2) whether the cost
function is minimized or not.

5. Explain Radial basis functions?

Ans -A Radial Basis Function (RBF), also known as kernel function, is applied to the distance to calculate
every neuron's weight (influence). The name of the Radial Basis Function comes from the radius distance,
which is the argument to the function.

6. What is Unfolding in time?

Ans - By unrolling we mean that we write out the network for the complete sequence. For example, if the
sequence we care about is a sentence of 3 words, the network would be unrolled into a 3-layer neural
network, one layer for each word. Input: x(t) is taken as the input to the network at time step t.

7. What is Teacher forcing?

Ans - Teacher forcing is an algorithm for training the weights of recurrent neural networks (RNNs). It
involves feeding observed sequence values (i.e. ground-truth samples) back into the RNN after each step,
thus forcing the RNN to stay close to the ground-truth sequence.

8. Explain Adaptation of weights?

Ans - learning means to do and adapt the change in itself as and when there is a change in environment. ANN
is a complex system or more precisely we can say that it is a complex adaptive system, which can change its
internal structure based on the information passing through it.

Long Question
1. Explain the limitations of backpropagation learning. Also explain the scope to overcome these
limitations?
2. What is radial basis function in neural network, explain in details?
3. What are dierent types of learning schemes used in training of artificial neural networks? Explain
each of them clearly with suitable examples?
4. Bring out the limitations of single layer Perceptrons in computing the logical functions. Suggest an
alternative network to overcome the above limitations and explain the suitable learning rule for the
 suggested network?
5. Bring out the limitations of single layer Perceptrons in computing the logical functions. Suggest an
alternative network to overcome the above limitations and explain the suitable learning rule for the
suggested network.
6. Explain the Evolutionary algorithms?
7. Explain Resilient Back propagation?
8. How the information is processed in Information processing of an RBF network?
9. Define the problem of handwritten digit recognition. With suitable diagram, explain architecture of
multilayer feed forward network for handwritten character recognition?
10. Calculate the total error of the network after using identity binary and bipolar activation function?
 Unit IV

Short Question
1. How can neural network be unsupervised?
Ans - Unsupervised learning means you're only exposing a machine to input data. There is no corresponding
output data to teach the system the answers it should be arriving at. With unsupervised learning, you train the
machine with unlabeled data that offers it no hints about what it's seeing.

2. Explain maxnet?
Ans - The maxnet is a fully connected network with each node connecting to every other nodes, including
itself. The basic idea is that the nodes compete against each other by sending out inhibiting signals to each
other. Using the above algorithm, all nodes converge to 0 except for the node with the maximum initial
value.

3. Explain hamming network?

Ans - Hamming neural network method is to decide which image of the reference vector closest to the input
vector. The decision in the Hamming neural network is taken recurrently layer. In recurrently layer is present
on one neuron for each reference image.

4. Write the algorithm of hamming network?

Ans -

5. What do you mean by Functionality and Training?

Ans - A neural network is a method in artificial intelligence that teaches computers to process data in a way
that is inspired by the human brain. It is a type of machine learning process, called deep learning, that uses
interconnected nodes or neurons in a layered structure that resembles the human brain.

Training an ANN is an iterative process in which training data examples are presented to the network one by
one, and the values of the weights are adjusted each time. After all examples get run through the network,
one training epoch is finished and the process often starts again.

6. What are the different topological functions?

Ans - There are two Artificial Neural Network topologies − FeedForward and Feedback.

7. Explain neural gas?

Ans - Neural Gas network is a competitive Artificial Neural Network (ANN), very similar to Self-Organizing
Map (SOM), which is proposed by Martinetz and Schulten, 1991. Neural Gas network can be used to solve
unsupervised learning tasks, like clustering, dimensionality reduction, and topology learning.

8. Explain multi neural gas?

Ans - Neural Gas network is a competitive Artificial Neural Network (ANN), very similar to Self-Organizing
Map (SOM), which is proposed by Martinetz and Schulten, 1991. Neural Gas network can be used to solve
unsupervised learning tasks, like clustering, dimensionality reduction, and topology learning. It has many
applications in the fields of pattern recognition, data compression, speech recognition, and image
segmentation.

9. What is growing neural gas?

Ans - The growing neural gas (GNG) is an unsupervised topology learning algorithm that models a data
space through interconnected units that stand on the most populated areas of that space. Its output is a graph
that can be visually represented on a two-dimensional plane, disclosing cluster patterns in datasets.

Long Question
1. Why reset mechanism is essential in ART network? Explain
2. Explain training algorithm of ART Network.
3. Explain the architectures of popular self-organizing maps. Derive the training algorithm of Kohonen
network.
4. Explain Kohonen’s self-organized feature map algorithm and mention its applications.
5. Consider a kohenen net with two cluster (output) units and five input units. The weight vector for the
output unit are W1=[1, 0.8,0.6,0.4,0.2] and W2=[1, 0.5,1,0.5,1]. Use the square of Euclidean distance to
find the winning neuron for the input pattern X=[0.5,1,0.5,0,0.5]. find the new weight for the winning
unit. Assume learning rate as 0.2?
6. Construct a maxnet with four neurons and inhibitory weight ᵋ =0.4, given the initial activation (input
signals) as follows a1(0)=0.2, a2(0)=0.6, a3(0)=0.5, a4(0)=0.4?
7. Briefly discuss about Hamming Net with example?
8. Construct and test the hamming network to cluster four vectors, given the exemplar vectors
e(1)=[1, -1, -1, -1]; e(2)=[-1, -1, -1, 1] the bipolar vectors are X1=[-1, -1,1, -1]; X2=[-1, -1,1, 1]; X3=[-1,
-1,-1, 1]; X4=[1, -1,-1, -1]
9. Construct an ART1 network for clustering four input vectors with low vigilance parameters of 0.4 into
three clusters. The four input vectors are [0001], [0101], [0011] and [1000]. Assume the necessary
parameters needed?
10. Construct an ART1 network for clustering four input vectors with low vigilance parameters of 0.4 into
three clusters. The four input vectors are [1001], [1101], [1011] and [1001]. Assume the necessary
parameters needed?