BANSAL INSTITUTE OF ENGINEERING AND
TECHNOLOGY
Department of CSE(AI)
Model Paper/ University Paper / Question Bank
Deep Learning (BAI-701)
Unit 1
2 Mark Questions
[Link]. Questions Session CO
a. What is Machine Learning ? 2025- 1
26
b. What are linear models? 2025- 1
26
c. Describe Perceptron 2025- 1
26
d. What are weights and Biases? 2025- 1
26
e. List types of Machine Learning. 2025- 1
26
f. Give 3 difference between Machine Learning and Deep Learning. 2025- 1
26
g. Define Deep Learning. 2025- 1
26
7 Mark Questions
[Link]. Questions Sessio CO
n
a. Describe how perceptron works? What are its limitations. 2025- 1
26
b. How does SVM determine the optimal Decision Boundary. 2025- 1
26
c. Explain the architecture of a back propagation network. What are the 2025- 1
limitations of back propagation algorithm? 26
d. Explain the architecture and functioning of a perceptron network. 2025- 1
26
e. List the various types of activation functions used in neural networks. 2025- 1
26
f. What do you understand by linearly separable problems? Give 2025- 1
examples. 26
g. Compare single layer and multilayer neural networks. 2025- 1
26
Unit 2
2 Mark Questions
[Link]. Questions Session CO
a. Write a short note on GANS 2025- 2
26
b. Write about batch normalization. 2025- 2
26
c. Write a short note regularization. 2025- 2
26
BANSAL INSTITUTE OF ENGINEERING AND
TECHNOLOGY
Department of CSE(AI)
d. What is overfitting and underfitting 2025- 2
26
e. What is Dropout 2025- 2
26
f. What is RELU? 2025- 2
26
g. List two advantages of Deeper Neural Network. 2025- 2
26
7 Mark Questions
[Link]. Questions Sessio CO
n
a. Difference between Deep and Shallow Network 2025- 2
26
b. Why CNN is preferred over ANN for Image Classification tasks even 2025- 2
though it is possible to solve image classification using ANN? 26
c. Explain GAN and its models. Name and describe different types of 2025- 2
GANs. 26
d. Examine the Semi –Supervised learning 2025- 2
26
e. Explain in detail about Regularization and its type. 2025- 2
26
f. Write about the significance of AlexNet I the evolution of Deep 2025- 2
Learning. 26
g. Explain how Dropout works with an example. 2025- 2
26
Unit 3
2 Mark Questions
[Link]. Questions Session CO
a. Why do we use convolutions for images instead of using fully 2025- 3
connected layers? 26
b. Define the different layers of a convolutional neural network. 2025- 3
26
c. Which is the best algorithm for face detection ? 2025- 3
26
d. What is autoencoder? 2025- 3
26
e. What is curse of Dimensionality? 2025- 3
26
f. Differentiate between PCA and LDA 2025- 3
26
g. What is LDA? 2025- 3
26
7 Mark Questions
[Link]. Questions Sessio CO
n
a. Draw and explain the architecture of Convolutional Networks. 2025- 3
26
BANSAL INSTITUTE OF ENGINEERING AND
TECHNOLOGY
Department of CSE(AI)
b. Explain in detail about VGG. 2025- 3
26
c. Compare PCA (Principle Component Analysis ) and RNN 2025- 3
26
d. Explain in detail about working of Autoencoder Architecture. 2025- 3
26
e. How dimensionality Reduction improves performance of Deep 2025- 3
Learning Models 26
f. Comapre and contrast traditional PCA and Deep learning Based 2025- 3
Autoencoders. 26
g. Discuss the role of Eigenvalues and Eigenvectors. 2025- 3
26
Unit 4
2 Mark Questions
[Link]. Questions Session CO
a. Explain in detail about Deep Reinforcement 2025- 4
26
b. What is RNN? 2025- 4
26
c. What is LSTM 2025- 4
26
d. Define Optimization in Neural Network. 2025- 4
26
e. What is ANN? 2025- 4
26
f. What is Deep Q networks? 2025- 4
26
g. Write a short note on GRU. 2025- 4
26
7 Mark Questions
[Link]. Questions Sessio CO
n
a. How AI and Neuroscience drive each other forwards? Explain. 2025- 4
26
b. Explain in detail about Deep Reinforcement 2025- 4
26
c. Explain LSTM (Long Short-Term Memory). Give some famous 2025- 4
applications of LSTM. 26
d. Differentiate between RNN, LSTM, and GRU 2025- 4
26
e. Discuss how the combination of dropout, batch normalization, and 2025- 4
early stopping together to improves model performance. 26
f. What role does dropout play in improving generalizations? How does 2025- 4
early stopping help in avoiding overfitting? 26
g. Explain working of Deep Q networks with diagram. 2025- 4
26
Unit 5
BANSAL INSTITUTE OF ENGINEERING AND
TECHNOLOGY
Department of CSE(AI)
2 Mark Questions
[Link]. Questions Session CO
h. What is Word2Vec in NLP? 2025- 5
26
i. What is Primary task of object detection? 2025- 5
26
j. Mention One application of WaveNet in audio processing? 2025- 5
26
k. What is YOLO? 2025- 5
26
l. What is NLP? 2025- 5
26
m. What is Alex Net? 2025- 5
26
n. What is RESNET used for ? 2025- 5
26
7 Mark Questions
[Link]. Questions Sessio CO
n
h. How does YOLO differs from traditional object detection pipeline? 2025- 5
26
i. Demonstrate how the privacy will be affected when facial recognition 2025- 5
gets used by private companies? 26
j. What role CNN play in bioinformatics joint detection? 2025- 5
26
k. How can an image gathering using AI enhance the performance of 2025- 5
training datasets? 26
l. Explain how context modelling in deep learning helps understand 2025- 5
complex scenes. 26
m. What are the ethical concerns surrounding face recognition 2025- 5
technologies? 26
n. Describe a case study where scene understanding improved safety or 2025- 5
decision making? 26