TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
Subject Name : DEEP LEARNING
Subject Code : 20AI7PC17
Degree & Branch : B.Tech – AIML
Regulation : R20
Year &Sem : IV - B.Tech & I-Sem
Faculty Name : J. RACHANA
PART –A
BLOOMS
S.No UNIT - I MARKS CO’s LEVELS
1
Define Deep Learning? 2.5 CO1 Remember
Explain advantages and disadvantages of
2 Deep Learning? 2.5 CO2 Understanding
Illustrate the Applications of Deep Learning?
3 2.5 CO3 Apply
Differentiate between Machine learning and
4 2.5 CO4 Analysing
Deep Learning
5 Discuss about Bayesian Learning 2.5 CO2 Understanding
Differentiate between Feed farward NN and
6 2.5 CO4 Analysing
Back propagation
7 What are Sigmoid neurons 2.5 CO1 Remember
PART –B
BLOOMS
S.No UNIT - I MARKS CO’s LEVELS
Define Deep Learning. Briefly explain History
a 5 CO4 Analysing
and success stories of Deep learning.
1 Elaborate the structure of basic Artificial
b Neuron(mc-culloch pitts model) with their 5 CO2 Understanding
functioning
Explain about Decision surfaces and Linear 10 CO4 Analysing
2
Classifiers with an example
TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
Illustrate with an example abou Linear 10
3 machines with Hinge Loss in DL CO3 Apply
Explain about Optimization techniques and 10
4 CO2 Understanding
Gradient Descent in deatail
5 Explain back-propagation algorithm in detail. 10 CO2 Understanding
PART –A
BLOOMS
S.No UNIT - II MARKS CO’s LEVELS
1
What is perceptron? 2.5 CO1 Remember
Differentiate between supervised and CO4 Analysing
2 2.5
unsupervised Learning
3 What is Exploding Gradient Problem in ANN 2.5 CO3 Apply
How can u choose Loss function in Deep
4 2.5 CO3 Apply
learning
5 What is Vanishning Gradient Problem in ANN 2.5 CO3 Apply
What type of problems can be solved by single
6 2.5 CO3 Apply
layer perceptrons
7 What is adaptive filtering problem 2.5 CO1 Remember
8 Differentiate between adaline and perceptron 2.5 CO4 Analysing
What is the relation between perceptron and CO4 Analysing
9 2.5
bayes classifier for a Guassian environ ment
10 What are learning rate Anealing techniques 2.5 CO1 Remember
PART –B
BLOOMS
S.No UNIT - II MARKS CO’s LEVELS
Write algorithm for Stochastic Gradient
a 5 CO5 evaluate
Descent
1 Explain single layer continuous training
b algorithm for solving linearly separable 5 CO3 Apply
problems.
a In which manner multilayer perceptron differs 5 CO4 Analysing
from single layer perceptron model?Explain
2 the reasons for emerging of multi-layer
perceptron.
b Write in detail about Unconstrained 5 CO2 Understanding
TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
organization techniques and Learning curves
a Discuss how Auto encoders are related to PCA 5 CO2 Understanding
3 b Differentiate between Auto encoder and 5 CO4 Analysing
Variational Auto encoder
Define heuristics Explain about output 10 CO4 Analysing
4
representation and Decision rule
Explain in detail how auto encoders are used 10
5 CO2 Understanding
in feature Generation.
PART –A
BLOOMS
S.No UNIT - III MARKS CO’s LEVELS
How would you define “Transfer
Learning”? List some transfer learning
1
models? 2.5 CO1 Remember
What are the differences between
pretraining and finetuning?
2 2.5 CO1 Remember
What do you mean by Convolutional Neural
3 Network 2.5 CO2 Understanding
CO4 Analysing
Why do we prefer Convolutional Neural
4 networks (CNN) over Artificial Neural 2.5
networks (ANN) for image data as input
Discuss about DeepNet-early stopping and
5 2.5 CO1 Remember
Dropout in CNN.
PART –B
BLOOMS
S.No UNIT - III MARKS CO’s LEVELS
Explain about Momentum Optimizer, RMs
1 5 CO2 Understanding
Prop, Adam in detail
TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
a 5 CO4 Analysing
Explain the significance of the RELU
2 Activation function in Convolution Neural
Network.
b List and explain virtues and limitations of 5
CO2 Understanding
CNN
3 Explain in Detail about CNN with example 10 CO4 Analysing
a What are the Different ways to Transfer 5 CO4 Analysing
Learning?
4
b Write about pros and cons of Transfer 5 CO4 Analysing
Learning
Explain about Dropout and Early atopping CO4 Analysing
5 10
Regularization in Deep nets
PART –A
BLOOMS
S.No UNIT - IV MARKS CO’s LEVELS
CO4 Analysing
1 Why do we use a Pooling Layer in a
CNN? 2.5
What is the size of the feature map for a
2 given input size image, Filter Size, Stride, 2.5 CO1 Remember
and Padding amount?
3 What is Stride? What is the effect of high 2.5 CO1 Remember
Stride on the feature map?
What is Word Embedding
4 2.5 CO1 Remember
What are supervised learning techniques in
5 2.5 CO1 Remember
Deep Learning
TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
PART –B
BLOOMS
S.No UNIT - IV MARKS CO’s LEVELS
Write a short notes on Approaches for Text CO4 Analysing
1 a 5
Representation in DL
a Write Advantages and Disadvantages of Word 5 CO4 Analysing
2
Embeddings
b Illustrate Continuous Bag of Words &
Skip-Gram
o
Explain learning vector quantization in detail 10 CO4 Analysing
3 with example
Write short notes on Supervised Learning 10 CO4 Analysing
4
models in Deep Learning.
5 Differentiate between ANN,RNN,CNN 10 CO4 Analysing
PART –A
BLOOMS
S.No UNIT - V MARKS CO’s LEVELS
1 What are the main concepts of encoder
2.5 CO1 Remember
decoder in Deep Learning
List applications of generative adversarial
2 2.5 CO2 Understanding
models
CO4 Analysing
3 What is an Attention Mechanism? 2.5
Advantages of Attention Mechanism in Deep
4 Learning Models 2.5 CO1 Remember
How does a generative model differ from a
5 discriminative model? 2.5 CO2 Understanding
TEEGALA KRISHNA REDDY ENGINEERING COLLEGE
(UGC-Autonomous)
(Sponsored by TKR Educational Society, Approved by AICTE, Affiliated to JNTUH
Accredited by NAAC with ‘A’ Grade. Accredited by NBA
Medbowli, Meerpet, Balapur(M), Hyderabad, Telangana- 500097
Mob: 8498085218. Email:
[email protected], www.tkrec.ac.in
PART –B
BLOOMS
S.No UNIT - V MARKS CO’s LEVELS
a What are Generative adversarial networks 5 CO5 Evaluate
1
b How do variational Encoders Work 5 CO4 Analysing
a 5
What are Applications and their Consequences
of Deep Generative models CO2 Understanding
b 5
2
Draw the architecture of encoder decoder in
CO2 Understanding
DL? How the Sequence to Sequence Model
works
a 5 CO4 Analysing
How Does the Attention Mechanism Work?
3
b How do attention mechanisms improve 5 CO4 Analysing
encoder decoder problems
10 CO4 Analysing
What are Basic Concepts of the Attention
4 Mechanism in Deep Learning Models
With the help of example Explain the concept 10 CO4 Analysing
5
of image classification using attention models
