COMPUTER SCIENCE & ENGINEERING-AI A &B DEPARTMENT

QUESTION BANK
DEEP LEARNING III year I sem
[Link] Questions marks CO PO BTL
1 What is the primary function of the thresholding logic in a McCulloch- 2M 1 1 1
Pitts unit?
2 Explain briefly define the concept of a "computational unit" in the 2M 1 1 2
context of artificial neural networks.
3 What is the key difference between a linearly separable and a non- 2M 1 2 1
linearly separable dataset?
4 when does a weight update occur, in the Perceptron Learning 2M 1 1 1
Algorithm?
5 What is the significance of the Perceptron Convergence Theorem? 2M 1 2 1
6 List two essential components of a Linear Perceptron. 2M 1 1 4
7 What is the primary function of a hidden layer in a Multilayer 2M 2 1 1
Perceptron?
8 Explain briefly define the role of Gradient Descent in training neural 2M 2 2 2
networks.
9 What problem does Backpropagation primarily solve? 2M 2 2 1
10 Name two common regularization techniques used in neural networks. 2M 2 5 1
11 What is the main objective of an Autoencoder? 2M 2 5 1
12 What is the core idea behind greedy layer-wise training? 2M 2 5 1
13 What is the main idea behind the Adagrad optimization algorithm? 2M 3 5 1
14 How does RMSProp handle the learning rate differently from Adagrad? 2M 3 2 1
15 What is the purpose of using dropout in neural networks? 2M 3 2 1
16 Define the saddle point problem in the context of neural network 2M 3 1 1
training.
17 What is the key advantage of using batch normalization? 2M 3 3 1
18 Name one difference between first-order and second-order optimization 2M 3 2 1
methods.
19 What is the purpose of Backpropagation Through Time (BPTT) in 2M 4 5 1
RNNs?
20 Name the three gates used in an LSTM cell. 2M 4 1 1
21 What is the main difference between a GRU and an LSTM? 2M 4 2 1
22 Why are Bidirectional RNNs useful in sequence modeling tasks? 2M 4 1 1
23 Who proposed the LeNet architecture and for what task was it 2M 4 1 2
originally designed?
24 List two major innovations introduced by AlexNet. 2M 4 4 4
25 What is the primary function of Variational Autoencoders (VAEs) in 2M 5 3 1
deep learning?
26 What is the key innovation of Transformers in natural language 2M 5 1 1
processing?
27 What is the primary application of GPT models in natural language 2M 5 5 1
processing?
28 How do VAEs differ from traditional autoencoders? 2M 5 2 2
29 What role do attention mechanisms play in Transformers? 2M 5 1 1
30 What are some applications of Transformers in computer vision? 2M 5 5 5

UNIT-I
[Link] Question Marks CO PO BTL
1 Explain the McCulloch-Pitts (MCP) neuron model, detailing its 7M 1 1 2
components and how it operates with thresholding logic.
2 Describe the Perceptron Learning Algorithm (PLA). Outline the steps 7M 1 2 I
involved in its iterative learning process, including the weight update
rule.
3 Define linear separability in the context of classification. Illustrate 7M 1 2 I
with examples of both linearly separable and non-linearly separable
datasets.
4 State and explain the Perceptron Convergence Theorem. What are 7M 1 2 2
the conditions under which the Perceptron Learning Algorithm is
guaranteed to converge?
5 Compare and contrast the structure and function of a biological 7M 1 1 2
neuron with that of an artificial computational unit like the Linear
Perceptron.
6 Discuss introduce the concept of Image Classification. How do the 7M 1 2 6
foundational ideas of computational units and linear separability
relate to basic approaches in image classification?

UNIT-II
[Link] Question Marks CO PO BTL
1 Explain the architecture of a Multilayer Perceptron (MLP), detailing 7M 2 1 5
the roles of its different layers (input, hidden, output) and how they
contribute to its functionality.
2 Explain the Backpropagation algorithm in detail, explaining how it 7M 2 1 2
efficiently computes gradients for weight updates in a feedforward
neural network.
3 Discuss the problem of vanishing/exploding gradients in deep neural 7M 2 2 6
networks and explain how greedy layer-wise pre-training can mitigate
this issue.
4 Explain the concept of regularization in neural networks. Describe at 7M 2 2 2
least two common regularization techniques (e.g., L1/L2
regularization, dropout) and their benefits.
5 What is an Autoencoder? Describe its architecture and explain its 7M 2 1 1
primary purpose, providing an example of its application.
 6 Explain the process of Gradient Descent as an optimization algorithm 7M 2 1 2
in the context of training neural networks. Discuss its role in
minimizing the loss function.
UNIT-III
[Link] Question Marks CO PO BTL
1 Explain the working of the Adam optimizer and compare it with 7M 3 1 2
RMSProp and Adagrad.
2 Discuss the significance of second-order optimization methods in 7M 3 2 6
training neural networks. How do they differ from first-order
methods?
3 What is the saddle point problem in neural network training? How 7M 5 1
3
does it affect the convergence of optimization algorithms?
4 Demonstrate the concept of Nesterov Accelerated Gradient (NAG) 7M 2 2
3
and how it improves over standard momentum-based methods.
5 What are the key differences between dropout and drop connect 7M 3 5 1
regularization methods? How do they help prevent overfitting?
6 Explain the role of Batch Normalization in neural networks and 7M 2 5
3
discuss its impact on training speed and performance.

UNIT-IV
[Link] Question Marks CO PO BTL
1 Explain the Backpropagation Through Time (BPTT) algorithm 7M 4 1 5
used for training Recurrent Neural Networks.
2 Explain the architecture and working of Long Short-Term 7M 4 1 5
Memory (LSTM) networks. How do they overcome the
vanishing gradient problem?
3 Compare and contrast Gated Recurrent Units (GRUs) with 7M 4 2 2
LSTMs in terms of structure and performance.
4 What are Bidirectional RNNs and Bidirectional LSTMs? 7M 4 5 1
Discuss their advantages and applications.
5 Explain the architecture of LeNet and its significance in the 7M 4 1 5
development of Convolutional Neural Networks (CNNs).
6 Explain the key features and innovations introduced by AlexNet 7M 4 2 5
that led to its success in image classification.

UNIT-V
[Link] Question Marks CO PO BTL
1 Explain the Transformer architecture and its role in Natural 7M 5 1 5
Language Processing (NLP).
2 Demonstrate the application of Variational Autoencoders (VAEs) in 7M 5 2 2
image synthesis and their benefits.
3 What are the applications of Transformers in speech processing? 7M 5 5 1
Explain their advantages.
4 How can Transformers and GPT be used for Multimodal AI tasks? 7M 5 2 1
Provide examples.
5 What are the ethical challenges associated with large generative AI 7M 5 5 1
models? Suggest mitigation strategies.

6 Explain the generative capabilities of GPT models and their 7M 5 2 5