ARTIFICIAL NEURAL NETWORK (2019 PATTERN)
Q1/2 . Hopfield Neural Network
1. What is a Hopfield neural network? Explain its working and state its limitations.
2. What is a state transition diagram for the Hopfield neural network? Explain how to derive it in the
Hopfield model.
3. Discuss the applications of Hopfield networks, such as solving the Travelling Salesman Problem.
Q1/2 . Boltzmann Machine
1. Describe the architecture of the Boltzmann machine.
2. Explain the Boltzmann learning law and its limitations.
3. How does the Boltzmann machine differ from the Hopfield network?
4. Write a short note on the applications of the Boltzmann machine.
Q1/2 . Associative Learning and Memory
1. Explain the concept of associative learning in artificial neural networks and its relation to pattern
recognition.
2. What is associative memory? Mention its characteristics and applications.
3. Define the following:
○ Pattern association
○ Pattern classification
○ Pattern mapping tasks
Q3/4 . Competitive Learning and Self-Organizing Maps (SOM)
1. What is competitive learning in neural networks? Explain its working and limitations.
2. Draw and explain the architecture of a Competitive Learning Network.
3. Describe the self-organizing map (SOM) algorithm and its applications in feature mapping.
4. Why are Kohonen networks called self-organizing maps?
Q3/4 . Adaptive Resonance Theory (ART)
1. Explain the architecture and features of the ART network.
2. Describe the working and training process of an ART network.
3. How can ART be used for character recognition tasks?
APR,OCT2023 & JUNE2024
ARTIFICIAL NEURAL NETWORK (2019 PATTERN)
4. Consider an ART-I network with input vectors [1,1,0,0], [0,0,1,0], [1,1,1,0], and [1,1,1,1], and the
following parameters: n=4,m=3,ρ=0.5n=4, m=3, \rho=0.5n=4,m=3,ρ=0.5. Perform clustering for
the first iteration.
Q 3/4 . Vector Quantization and Clustering
1. What is vector quantization? How is it used for pattern clustering?
2. Define:
○ Learning vector quantization
○ Adaptive pattern classification
Q 5/6 . Convolutional Neural Networks (CNN)
1. What is a Convolutional Neural Network (CNN), and how does it work?
2. Why are CNNs preferred over Artificial Neural Networks (ANNs) for image data?
3. Explain the following components of CNN:
○ Pooling layer and its role
○ Flattening layer and its function
○ Padding in neural networks
4. Illustrate convolution and max pooling with an example.
5. Discuss advanced CNN architectures:
○ AlexNet
○ VGG-16
○ Residual Networks
Q 5/6 .Deep Learning Concepts
1. Explain transfer learning and its importance in deep learning.
2. What are the different types of transfer learning models for image classification?
3. Define bias and variance. What is the bias-variance tradeoff?
4. Write a short note on frameworks used in deep learning (mention seven).
Q 7/8 . Pattern Recognition and Applications
1. Explain how artificial neural networks (ANNs) can be used for recognizing printed and
handwritten characters.
2. Describe the Neocognitron model and its role in pattern recognition.
APR,OCT2023 & JUNE2024
ARTIFICIAL NEURAL NETWORK (2019 PATTERN)
3. Discuss an example of pattern recognition in everyday life.
4. Explain the recognition of Olympic Games symbols using ANN.
5. Discuss texture classification and segmentation using ANN.
Q1/2 . Advanced Topics in Neural Networks
1. What is simulated annealing? Explain its advantages and disadvantages.
2. How does the simulated annealing algorithm work?
3. Explain stochastic gradient descent and its applications.
4. What is catastrophic forgetting in neural networks?
Q 7/8 .Miscellaneous Topics
1. Explain automatic language translation, its working, and its benefits.
2. Write short notes on the following:
○ State transition diagram
○ False minima problem
○ NET Talk
APR,OCT2023 & JUNE2024
