Course Name : Deep Learning
Course Code : AIN 5002/DS2702/CS2702
LTP : 3 0 2
Course Objectives:
Students should be able to:
understand the fundamentals of deep learning
apply different deep learning models using various datasets
learn various architectures and optimization methods
learn the practical knowledge in handling and analysing real world applications and
recent research activities in this field.
Course Contents: Total No. of Lectures – 42
Lecture wise breakup No. of
Lectures
Unit 1 Linear Algebra Review and Optimization 6
Brief review of concepts from Linear Algebra, Types of errors,
bias-variance trade-off, overfitting-underfitting, brief review of
concepts from Vector Calculus and optimization, variants of
gradient descent, momentum
Unit 2 Logistic Regression 6
Basic concepts of regression and classification problems, linear
models addressing regression and classification, maximum
likelihood, logistic regression classifiers
Unit 3 Neural Networks 6
Basic concepts of artificial neurons, single and multi-layer
perceptron, perceptron learning algorithm, different activation
functions, cross entropy loss function.
Unit 4 ConvNets 7
Basic concepts of Convolutional Neural Networks, Convolution and
pooling operation, Transfer learning, convnet architectures -
AlexNet, ZFNet, VGG, GoogLeNet, ResNet
Unit 5 Recurrent Neural Networks 7
Basic concepts of Recurrent Neural Networks (RNNs),
backpropagation through time, Long-Short Term Memory (LSTM)
architectures, the problem of exploding and vanishing gradients,
and basics of word embedding
Unit 6 Regularization, Batchnorm 5
Discussion on regularization, L1-Regularization, L2-regularization,
Dropout, Batchnorm
Unit 7 Autoencoders: 5
Autoencoders, Denoising autoencoders, sparse autoencoders,
contractive autoencoders
List of Experiments: Number of
Turns
1 Introduction to python libraries for deep learning: Keras, Tensorflow, 2
OpenCV
2 Implementation of MultiLayer Perceptron(MLP) 2
3 Implementation Basic CNN model and various CNN architectures - 4
transfer learning
4 Implementation of Recurrent neural networks 2
5 Deep Learning based project 4
Course Outcomes
At the end of the course, students will be able to:
1 Understand the fundamentals of deep learning and its applications.
2 Identify and apply the learned concepts and methods to a real-world problem
3 Analyse, design and implement various deep learning algorithms
4 Evaluate and test the efficacy of the developed model.
Bibliography:
Text Book:
Sr. Name of Book/ Authors/ Publisher Year of
No. Publication
/ Reprint
1 “Deep Learning” by Ian Goodfellow, Yoshua Bengio, Aaron Courville, 2016
MIT press
Reference Books:
Sr. Name of Book/ Authors/ Publisher Year of
No. Publication
/ Reprint
1 “Learning deep architectures for AI: Foundations and trends in 2009
Machine Learning” by Yoshua Bengio, Now Publishers Inc.
2 “Deep Learning with TensorFlow: Explore neural networks with 2017
Python”, Giancarlo Zaccone, Md. RezaulKarim, Ahmed Menshawy,
Packt Publisher.
3 “Deep Learning from Scratch: Building with Python from First 2019
Principles”, Seth Weidman, O’Reilly
Equivalent MOOCs courses:
Sr. No. Course Links Offered by
1 https://archive.nptel.ac.in/courses/106/106/106106184/ NPTEL
2 https://onlinecourses.nptel.ac.in/noc20_cs62/preview NPTEL
Marks Distribution
*Assignments/Quiz 25% of assignments given in the
course/as given by PEC
Lab Evaluation/Lab Quiz 10%
Mid Term 25%
Final Exam (End Term) 40%
*Please note that assignments encompass all types (including quizzes(min 3), programming
tasks, and essay submissions) available in the specific week.
Course Instructor: Dr. Padmavati
