LNMIIT, Jaipur

Department of Computer Science & Engineering

Programme: Course Title: Course Code:

B. Tech. (CSE) Deep Learning CSE4121
Type of Course: Prerequisites: Total Contact Hours:
Program NIL 40
Elective
Year/Semester: Lecture Hrs/Week: Tutorial Hrs/Week: Practical Hrs/Week: Credits:
3/Even 3 0 0 3

Learning Objectives:

This course covers the most successful form of artificial intelligence, deep learning. This course
covers linear regression, logistic regression, deep neural networks, convolutional, and recurrent
neural networks. This course also covers various aspects of deep learning including overfitting,
underfitting, optimization methods, and regularization techniques such as L1, L2, and dropout.
Programming will also be an important component of this course, using preferably python.

Course outcomes (COs):

On completion of this course, the students will have the ability to: Bloom’s Level
CO-1 Analyze and Understand usefulness of data 2, 4, 5
CO-2 Apply deep learning techniques to real-world problems 3
CO-3 Critically analyze which architecture to use for a specific problem 4
CO-4 Design and implement deep learning algorithms 3, 6

Course Topics Lecture Hours

UNIT – I (Introduction) 4
4
1.1 Linear and logistic regression. Cost function for logistic regression. 4

UNIT – II Deep Neural Networks 9

2.1 Generalization of logistic regression to deep neural networks. Cost 9

functions. Optimization algorithms: Gradient descent, Stochastic gradient 9
descent, Momentum, RMSprop, Adam.

CSE Department, LNMIIT Jaipur Page | 1-3

 LNMIIT, Jaipur
Department of Computer Science & Engineering

UNIT – III Regularization Techniques 6

3.1 Underfitting and overfitting of neural networks: bias and variance. L1, 6
L2, and dropout regularization techniques, hyperparameter tuning. 6

UNIT-IV Deep Learning for Computer Vision 12

4.1 Basics of CNN: convolutions and pooling. Detailed understanding of

Alexnet, ResNet, VGG-16, VGG-19, and inception architectures. Their 12
12
implementations. Object recognition and face recognition.

UNIT-V Deep Learning for Natural Language Processing 9

5.1 Basics of RNN, LSTM, GRU, Bidirectional RNN, deep RNNs.

Representations of words as vectors. One hot encoding and word embeddings. 9
9
Learning word embeddings using word2vec, glove, and fasttext.

Textbook References:
Text Books:
1. I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning, Cambridge, MA: MIT Press, 2016.
2. "Deep Learning: A Visual Approach" Book by Andrew Glassner.
3. "Practical Natural Language Processing" by Sowmya Vajjala, Bodhisattwa Majumder, Anuj
Gupta, Harshit Surana.
4. "Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow" by Aurélien Geron.

Additional Resources (NPTEL, MIT Video Lectures, Web resources etc.):

1. Stanford CS230: Deep Learning

2. MIT 6.S191 Introduction to Deep Learning (January-March 2018)
3. Stanford CS224d: Deep Learning for Natural Language Processing (January-March, 2018)
4. Stanford CS231n: Convolutional Neural Networks for Visual Recognition (Spring 2018)
5. Coursera specialization on Deep Learning
6. Dive into Deep Learning book https://d2l.ai/index.html
7. "The Deep Learning Revolution" Book by Terry Sejnowski.
8. Introduction to Deep Learning Book (The MIT Press) by Eugene Charniak

CSE Department, LNMIIT Jaipur Page | 2-3

 LNMIIT, Jaipur
Department of Computer Science & Engineering

Evaluation Method
Item Weightage (%)
Assignments 10
Quizzes 20
Project 10
Midterm 20
Endterm 40

CO and PO Correlation Matrix

CO PO1 PO2 PO3 PO4 PO5 PO6 PO7 PO8 PO9 PO10 PO11 PO12 PSO1 PSO2 PSO3
CO1 1 1 1 3 3 3 2 2
CO2 2 2 3 3 2 1 3 3 3 3 3 3 3
CO3 3 3 3 3 3 3 3 2 2
CO4 3 3 3 3 2 1 3 3 3 3 3 3 3

Last Updated On: 04 January 2023

Updated By: Preety Singh, Indradeep Mastan

CSE Department, LNMIIT Jaipur Page | 3-3