UNIT I: Introduction to Deep Learning
🔹 2-Mark Questions (Q1 subparts):
What is Deep Learning?
What is the promising future of AI?
What are the advantages of neural networks?
How Deep Learning differs from Machine Learning?
The "deep" in Deep Learning – what does it mean?
What challenges could you face when applying neural networks?
List out various latest examples of NLP.
Distinguish between biological and artificial neurons.
What has deep learning achieved so far?
🔹 14-Mark Questions:
Q2a: How Deep Learning works? What are various applications of Deep
Learning?
Q2b: Differentiate between Artificial Intelligence and Deep Learning.
Q2a: How Deep Learning works? Explain briefly.
Q2b: Explain the concept of learning representation from data.
UNIT II: Neural Networks and Learning Representation
🔹 2-Mark Questions (Q1 subparts):
Define Linear Perceptron.
Define Neural Network and discuss its properties.
What is a model in deep learning?
What are loss functions and optimizers?
What is a feed-forward neural network?
What are sigmoid and tanh functions?
What is the delta rule?
🔹 14-Mark Questions:
Q3a: Define Neural Network and discuss its properties.
Q3b: Explain Activation Function in detail.
Q3a: What is Perceptron? Explain the working.
Q3b: List and explain the various activation functions used in modeling
artificial neurons.
Q3a: Explain the concept of deep feed-forward neural networks.
Q3b: Explain activation function used in modeling artificial neurons and its
suitability.
UNIT III: Training Neural Networks & TensorFlow
🔹 2-Mark Questions (Q1 subparts):
What is TensorFlow?
What is Gradient Descent?
What is the delta rule?
Define Backpropagation.
What are placeholder tensors and sessions in TensorFlow?
Explain how to install TensorFlow.
Determine the shape of output matrix for 19×19 image, padding 2, stride 2,
5x5 filter.
🔹 14-Mark Questions:
Q4a: Write Backpropagation Algorithm and explain.
Q4b: Explain Delta Rule in detail.
Q4a: Explain the process of Gradient Descent and its role in optimizing
weights.
Q4b: Discuss building Multilayer Model for MNIST in TensorFlow.
Q7a: Explain in detail the workflow for training and evaluating a deep learning
model.
Q7b: Explain how to create and manipulate TensorFlow variables and
demonstrate operations.
Q4a: What is a feed-forward network? Explain gradient-based learning.
Q4b: How to build a neural network using TensorFlow? How to visualize it?
Q7a: What are various steps required for installation of TensorFlow? Discuss
TensorFlow operations.
Q7b: Explain about Backpropagation algorithm for training a neural network.
UNIT IV: Deep Network Architectures (CNN, RNN, UPN)
🔹 2-Mark Questions (Q1 subparts):
What is an Unsupervised Pretrained Network?
Describe pooling layer in CNN.
Describe the shape of input data for a Recurrent Neural Network.
🔹 14-Mark Questions:
Q5a: Explain CNN architecture in detail.
Q5b: Write basic steps to load CNN on MNIST.
Q5a: Explain about Convolutional Neural Network Architecture.
Q5b: Discuss about Recurrent Neural Network (RNN).
Q5a: With the help of diagram, explain basic building blocks of CNN.
Q5b: Discuss Recursive Neural Networks with neat diagram.
Q6b: Explain use of deep learning in speech recognition.
Q7b: Explain the architecture of LSTM.
UNIT V: Deep Learning Applications
🔹 2-Mark Questions (Q1 subparts):
Give two applications of Deep Learning.
What are the top five use cases of TensorFlow?
Mention use of DL in Speech Recognition and NLP.
🔹 14-Mark Questions:
Q6a: Write about various implementations available for large-scale deep
learning.
Q6b: Explain the use of deep learning in NLP.
Q6a: What are various implementations available for large-scale deep
learning?
Q6b: Explain the use of deep learning in Speech Recognition.
