B.M.S.

COLLEGE OF ENGINEERING, BENGALURU-19

Autonomous Institute, Affiliated to VTU
DEPARTMENT OF CSE (Data Science)

Sem V
Course Title: Deep Learning
Course Code: 23DS5PCDLG Total Contact Hours: 50 hours
L-T-P: 4-0-1 Total Credits: 5

Unit Topics Hrs.

No.
1 Introduction to Artificial Neural Networks: From Biological to Artificial Neurons: 10
Biological Neurons, Logical Computations with Neurons, The Perceptron, The
Multilayer Perceptron and Backpropagation, Regression and Classification MLPs,
Implementing MLPs with Keras, Fine-tuning Neural Network Parameters
Introduction to Deep Learning: Challenges motivating Deep Learning, Historical
Trends in Deep Learning, Deep Feedforward Networks, Gradient-based Learning,
and Efficient Computation.
2 Convolution Neural Networks (CNN): The Architecture of the Visual Cortex, 10
Convolutional Layer, Pooling Layer, CNN Architectures- LeNet-5, AlexNet,
GoogLeNet, VGGNet, ResNet, Xception, Pre-trained Models for Transfer Learning,
Classification and Localization, Object Detection-Fully Convolutional Networks
(FCNs), You Only Look Once (YOLO), Semantic Segmentation, Convolution
Networks and the History of Deep Learning.

3 Training Deep Neural Networks: Vanishing/Exploding Gradients, Reusing 10

Pretrained Layers, Faster Optimizers, Avoiding Overfitting Through Regularization.
Recurrent Neural Networks (RNN): Recurrent Neurons and Layers, Training
RNNs, forecasting a Time Series, Handling Long Sequences-LSTM, GRU,
Bidirectional RNNs, Recursive Neural Networks, Attention Mechanisms.
4 Representation Learning using Autoencoders: Stacked Autoencoders, 10
Convolutional Autoencoders, Recurrent Autoencoders, Denoising Autoencoders,
Sparse Autoencoders, Applications of Autoencoders, Transfer Learning, and Domain
Adaptation.
Generative Adversarial Networks (GANs): Difficulties of training GANs, Deep
Convolutional GANs, Progressive Growing of GANs, StyleGANs
 B.M.S. COLLEGE OF ENGINEERING, BENGALURU-19
Autonomous Institute, Affiliated to VTU
DEPARTMENT OF CSE (Data Science)
5 Deep Generative Models: Boltzmann Machines, Restricted Boltzmann Machines, 10
Deep Belief Networks, Deep Boltzmann Machines, Boltzmann Machines for Real-
Valued Data
Applications: Large-Scale Deep Learning, Computer Vision, Speech Recognition,
Natural Language Processing, Other Applications.

Prescribed Text Book

Sl. No. Book Title Authors Edition Publisher Year
1. Deep Learning Ian Goodfellow, MIT Press 2016
Yoshua Bengio, Aaron
Courville

2. Hands-On Machine Aurelien Geron O’Reilly 2020

Learning with Scikit- Second
Learn, Keras &
TensorFlow
Reference Text Book
Sl. No. Book Title Authors Edition Publisher Year
1. Deep Learning with Amita Kapoor, Third Packt 2022
Tensor Flow and Keras Antonio Gulli, Sujit
Pal
2. Learning Deep Magnus Ekman First Addison-Wesley 2021
Learning: Theory and Professional
Practice of Neural
Networks, Computer
Vision, NLP, and
Transformers using
TensorFlow
 B.M.S. COLLEGE OF ENGINEERING, BENGALURU-19
Autonomous Institute, Affiliated to VTU
DEPARTMENT OF CSE (Data Science)
E-Book
Sl. Book Title Authors Edition Publisher Year URL
No.
1. Dive into Aston Zhang, First Cambridge 2023 https://d2l.ai/d2l-en.pdf
Deep Zachary C.Lipton, University
Learning Mu Li, Alexander Press
J. Smola
2. The Little François Fleuret - University 2024 https://fleuret.org/public/lbdl.
Book of of Geneva, pdf
Deep Switzerland
Learning

MOOC Course
Sl. No. Course name Course Year URL
Offered By
1. Deep SWAYAM 2024 https://onlinecourses.nptel.ac.in/noc24_cs114/prev
learning – iew
IIT Ropar
2. Neural Coursera 2024 https://www.coursera.org/learn/neural-networks-
Networks deep-learning?specialization=deep-learning
and Deep
Learning

Laboratory Plan

Sl. No. Lab Program

1 Write a program to implement XOR gates using Perceptron.

2
Design a deep NN, optimize the network with Gradient Descent, and optimize the

same with Stochastic gradient descent(SGD).

3
Classification of MNIST Dataset using CNN.

4
Implement Region-Based CNN for object detection.
 B.M.S. COLLEGE OF ENGINEERING, BENGALURU-19
Autonomous Institute, Affiliated to VTU
DEPARTMENT OF CSE (Data Science)

5
Implement RNN for handwriting digit recognition.

6
Implement Bidirectional RNNs for music generation.

7
Implement Bidirectional LSTM for sentiment analysis.

8
Implement Variational Autoencoders for image-denoising.

9
Implementation of a Restricted Boltzmann Machine (RBM) that demonstrates
stacking.

10
Implement Generative Adversarial Networks to generate realistic photographs.

Course Outcomes

At the end of the course, the student will be able to

CO1 Apply the fundamentals of deep learning algorithms for various complex problems
Analyze appropriate deep learning concepts based on the characteristics of the
CO2
problem and the data.
Design models using deep learning to solve real-world problems and present solution in a
CO3
team
CO4 Conduct practical experiments and implement Deep Learning Models for Applications

CO-PO-PSO mapping

PO1 PO2 PO3 PO4 PO5 PO6 PO7 PO8 PO9 PO10 PO11 PO12
CO1 3
CO2 2
CO3 3 1 1 1
CO4 2 2
 B.M.S. COLLEGE OF ENGINEERING, BENGALURU-19
Autonomous Institute, Affiliated to VTU
DEPARTMENT OF CSE (Data Science)
Proposed Assessment Plan (for 50 marks of CIE)

Assessment Tool No. of Assessments Marks

Internals Best of 2 20
Lab Test + Project 1 25
QUIZ/AAT 1 5
Total 50

SEE Exam Question paper format

Unit-1 Mandatory One Question to be asked for 20 Marks

Unit-2 Internal Choice One Question to be asked for 20 Marks
Unit-3 Internal Choice One Question to be asked for 20 Marks
Unit-4 Mandatory One Question to be asked for 20 Marks
Unit-5 Mandatory One Question to be asked for 20 Marks