Course Title: Neural Networks
Course Code: PCC-CSE-401-G
Semester: B.Tech. 7th Semester (CSE)
2024 Examination
Short Answer Questions [6 × 2.5 = 15 marks]
- Explain the following:
(a) Role of Dendrites and Axons in a Biological Neuron
(b) Applications of Neural Network
(c) Linear Separability
(d) Supervised Learning
(e) Generalized Delta Learning rule
(f) Hetero Associative Memory
Section-A [15 marks]
- Describe the structure of a biological neuron in detail and explain how its components contribute to signal processing and transmission. [15]
- What are activation functions in artificial neural networks? Explain their role and importance with examples of threshold, sigmoid and tanh functions. [15]
Section-B [15 marks]
- Describe the McCulloch-Pitts (MCP) model in detail. Discuss its architecture, working and the solution for OR functions. [15]
- Illustrate the concept of linear separability with the solution of the OR function. [15]
Section-C [15 marks]
- Describe the architecture of Backpropagation network. Discuss the training and testing processes in detail. [15]
- Explain competitive learning. Discuss its applications and how it differs from conventional learning methods. [15]
Section-D [15 marks]
- What is Associative Memory? How does associative memory handle errors in data, such as missing or mistaken inputs? Discuss the mechanisms with examples. [15]
- Explain Hebb model and implement logic 'AND' function using Hebb network. [15]
November—2023 Examination
Short Answer Questions [6 × 2.5 = 15 marks]
- (a) Differentiate between Feedforward and Feedback Networks
(b) What are Separability limitations in unsupervised Learning?
(c) Define Activation function
(d) Reinforcement Learning
(e) Unsupervised Learning
(f) Backpropagation Network
Section-A [15 marks]
- Describe the following:
(a) Auto Associative Memory
(b) Hetero Associative Memory - How biological neural is different from the artificial neural networks? Explain architecture of biological neural network.
Section-B [15 marks]
- Explain Gradient Descent algorithm.
- Explain the following: (a) Application of Artificial Neural Network (b) McCulloch Pitts Model
Section-C [15 marks]
- Explain storage and retrieval algorithm for associative memory.
- In which manner multilayer perceptron model differ from single layer Perceptron Model? Explain the reasons for emergence of multilayer perceptron model.
Section-D [15 marks]
- Explain Hebbian Model and implement logic 'AND' function using Hebb Network.
- What do you mean by Generalized Delta rule? Explain in detail.
December - 2022 Examination
Short Answer Questions
- (a) Discuss evolution of neural network.
(b) Explain the Competitive Learning.
(c) Obtain the output of neuron Y having three inputs (x₁ = 1, x₂ = 2, x₃ = 3) and weights (w₁ = 1, w₂ = 2, w₃ = 3) by using Threshold and Sigmoidal activation functions.
(d) Discuss the concept of Storage capacity in Associative Memory.
Section-A [15 marks]
- Explain the component of a Biological Neuron. Also focus on Biological neuron equivalencies to artificial neuron model.
- What is Activation Function? Why it is used? Give different types of activation function in detail.
Section-B [15 marks]
- What is perceptron? Also realize it for OR function for bipolar data.
- Explain Linear Separability Concept by taking a suitable example, also classify the output of OR function using it.
Section-C [15 marks]
- Derive Gradient Descent algorithm and compare it with generalized delta learning rule.
- What is Learning? Explain its different types.
Section-D [15 marks]
- Store the vector (1 1 -1 -1) in Auto Associative Network, and:
(a) Find the Weight Matrix
(b) Test the net with input vector
(c) Test with one mistake in input
(d) Test with one missing in input
(e) Test with two missing in input
(f) Test with two mistakes in input - What is Associative Memory? Explain Auto Associative Memory with its architecture, training (insertion), and testing (retrieval) algorithm.