Master of Engineering - ME (Artificial Intelligence and

Machine Learning)
Syllabus

July 2021 Onwards

MANIPAL SCHOOL OF INFORMATION SCIENCES

MANIPAL ACADEMY OF HIGHER EDUCATION
MANIPAL – 576104, KARNATAKA, INDIA.
 Program Educational Objectives / Outcomes (PEOs)

PEO 1: Produce industry-ready graduates with solid foundation in fundamentals of machine learning
and practical experience in structuring machine learning projects using state of the art software.

PEO 2: Machine learning researchers who can innovate and address research challenges through
doctoral studies and professional roles in public/private research labs.

PEO 3: Entrepreneurial engineers who can identify and address real-life problems in sustainability,
environment, education, and governance.

Program Objectives / Outcomes (POs)

PO1 Scholarship of Knowledge: Acquire solid mathematical and computational skills essential for
understanding, applying, and developing modern machine learning algorithms.

PO2 Critical Thinking: Identify, formulate, analyse, and solve real-life problems using machine
learning principles.

PO3 Problem Solving: Identify appropriate and efficient algorithmic approaches for solving real-life
problems using machine learning principles.

PO4 Research Skill: Keep updated with current research trends in machine learning and innovate
research ideas for developing new machine learning paradigms.

PO5 Usage of modern tools: Gain solid skills in using state of the art modern machine learning
software prevalent in industry and academia.

PO6 Collaborative and Multidisciplinary work: Use machine learning as a common solution platform
to identify problems and collaborate with researchers from health care, natural & social sciences,
arts, and humanities.

PO7 Project Management and Finance: Streamline and realize project ideas into entrepreneurial
ventures involving good project management practices and financial considerations.

ME (Artificial Intelligence and Machine Learning) 1

 PO8 Communication: Professionally communicate the results of applying machine learning algorithms
to real life problems to aid decision making processes.

PO9 Life-long Learning: Evolve and adapt to the fast-changing artificial intelligence landscape
through academic and industrial engagements.

PO10 Ethical Practices and Social Responsibility: Through ethical practices, teamwork, and
leadership skills, use machine learning skills to address problems of social importance for sustainable
societal development.

PO11 Independent and Reflective Learning: Critically examine data and the interpretation of
outcomes of machine learning algorithms and take corrective measures without depending on external
feedback.

ME (Artificial Intelligence and Machine Learning) 2

 Program Structure
ME (Artificial Intelligence and Machine Learning) – I Semester

Duration of Exam
No. of Hrs./week Maximum Marks

Course

External 50
in Hrs

Internal 50
Course Name

Total 100
Practical
Tutorial
Lecture
Code

Credit
BDA 5101 Algorithms and Data
3 - - 3 3 50 50 100
Structures for Big Data
AML 5101 Applied Linear Algebra 3 - - 3 3 50 50 100

AML 5102 Applied Machine Learning 3 - - 3 3 50 50 100

AML 5103 Applied Probability and

3 - - 3 3 50 50 100
Statistics
Elective – I 3 - - 3 3 50 50 100

BDA 5151 Algorithms and Data

- - 3 1 3 50 50 100
Structures for Big Data Lab
AML 5151 Applied Linear Algebra Lab - - 3 1 3 50 50 100

AML 5152 Applied Machine Learning

- - 3 1 3 50 50 100
Lab
Applied Probability and
AML 5153 - - 3 1 3 50 50 100
Statistics Lab

Elective – I Lab - - 3 1 3 50 50 100

MPT 5100 Mini Project – I - - - 4 - 10 - 100
Professional Skill 0
10
PSD 5100 - - - 1 - - 100
Development – I 0

Total 15 - 15 25

ME (Artificial Intelligence and Machine Learning) 3

 ME (Artificial Intelligence and Machine Learning) – II Semester

Duration of Exam
No. of Hrs. / week Maximum Marks

Course

External 50
in Hrs

Internal 50
Course Name

Total 100
Practical
Tutorial
Lecture
Code

Credit
AML 5201 Advanced Applications of 3 - - 3 3 50 50 100
Probability & Statistics

AML 5202 Deep Learning 3 - - 3 3 50 50 100

AML 5203 Machine Learning 3 - - 3 3 50 50 100

Principles & Applications
AML 5204 Reinforcement Learning 3 - - 3 3 50 50 100

Elective – II 3 - - 3 3 50 50 100

AML 5251 Advanced Applications of - - 3 1 3 50 50 100

Probability & Statistics Lab
AML 5252 Deep Learning Lab - - 3 1 3 50 50 100

Machine Learning
AML 5253 Principles & Applications - - 3 1 3 50 50 100

Lab

AML 5254 Reinforcement Learning - - 3 1 3 50 50 100

Lab
Elective – II Lab - - 3 1 3 50 50 100
MPT 5200 Mini Project – II - - - 4 - 100 - 100

PSD 5200 Professional Skill - - - 1 - 100 - 100

Development – II
TOTAL 15 - 15 25

ME (Artificial Intelligence and Machine Learning) - III & IV Semesters

AML 6098 Project Work - - - 25

Total Number of Credits to Award Degree 75

ME (Artificial Intelligence and Machine Learning) 4

 List of Electives (Theory)

Elective - I Elective - II

Course Code Course Name Course Code Course Name

Applied Mathematics for

AML 5131 Applications of Graph Theory AML 5231
Machine Learning

BDA 5102 AML 5232 Convolutional Neural Networks

Architecture of Big Data Systems
for Computer Vision
BDA 5132 AML 5233 Natural Language Processing
Principles of Data Visualization
Principles & Applications
ENP 5230
Entrepreneurship

List of Electives (Lab)

Elective - I Elective - II

Course Code Course Name Course Code Course Name

Applied Mathematics for

AML 5181 Applications of Graph Theory Lab AML 5281
Machine Learning Lab

BDA 5152 Architecture of Big Data Systems AML 5282 Convolutional Neural Networks
Lab for Computer Vision Lab
BDA 5182 AML 5283 Natural Language Processing
Principles of Data Visualization Lab
Principles & Applications Lab
ENP 5280
Entrepreneurship Lab

ME (Artificial Intelligence and Machine Learning) 5

 SEMESTER I

L T P C Total hours
BDA 5101: Algorithms and Data Structures for Big Data
3 0 0 3 36

Course Outcome

1. Analyse recursive programs, solve a general class of recurrence relations.

2. Design programs for implementation of linked lists, stack, queues, binary search tree, sorting and
searching.
3. Design programs for dictionary, hash tables, graphs and shortest path techniques.
4. Design string and text processing programs.
No. of
Unit
Topics Hours

Algorithm specification and analysis techniques:

I 3
Analysis of recursive programs. Solving recurrence equations. General solution for a
large class of recurrences.

II Elementary data structures: 10

Implementation of lists, stacks, queues, Trees

Sorting & Searching, Hash Tables, Graph:

Quick sort, heap sort, merge sort. Linear search and binary search.
III 14
Hashing and Dictionaries
Representation of graphs. Depth First Searching. Breadth First Searching.
Minimum cost spanning tree. Single source shortest paths and all-pairs shortest path
String and text processing techniques, Data stream algorithms:
9
IV Pattern-Matching Algorithms. Text Compression. Tries
Sampling, Random Projections, Basic Algorithmic Techniques. Group Testing, Tree
Method and Graph sketching.
References

1. Introduction to Algorithms - Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest. MIT Press.
2. Data Structures and Algorithms - Aho, Hopcroft and Ulmann. Pearson Publishers.
3. Data Structures and Algorithms in Python - Michael T. Goodrich, Roberto Tamassia, and Michael H.
Goldwasser. John Wiley & Sons.

ME (Artificial Intelligence and Machine Learning) 6

4. Data Streams: Algorithms and Applications - S. Muthukrishnan. Foundations and Trends in Theoretical
Computer Science archive, Volume 1 Issue 2, August 2005, Pages 117 – 236

Total
L T P C
hours
AML 5101: Applied Linear Algebra

3 0 0 3 36

Course Outcome

1. Develop a solid understanding of matrix-vector operations and relate them to real-life calculations.
2. Apply and analyse algorithms constructed using matrix-vector principles.

3. Develop models for real-life applications using the least squares technique and interpret the results from a
practical perspective.
No. of
Unit Topics Hours

Vectors:
Conceptual introduction to vectors; vector addition; scalar-vector multiplication - Dot
product; norm; distance - Standard deviation; standardization vs. normalization;
I 10
angle between vectors - Application example: k-means clustering algorithm - Linear
dependence/independence; basis - Orthonormal vectors; projections; Gram-Schmidt
algorithm.
Matrices:
Conceptual introduction to matrices; types of matrices (zero, identity, diagonal) -
II Addition of matrices; transpose; norm - Matrix-vector product – concept & examples 14
- Systems of linear equations: over- & under-determined systems - Matrix-matrix
product – concept & examples - QR factorization - Solving linear equations.
Linear Least Squares:
Least squares: problem motivation and examples - Solving linear least squares
III problems - Least squares data fitting; validation; feature engineering - Least squares 12

classification.

References

1. Introduction to Applied Linear Algebra, Vectors, Matrices, and Least Squares, Stephen Boyd & Lieven
Vandenberghe, Cambridge University Press, 1st Edition, 2018. Available online at http://vmls-
book.stanford.edu/vmls.pdf

ME (Artificial Intelligence and Machine Learning) 7

2. Linear Algebra and Learning from Data, Gilbert Strang, Cambridge Uni. Press; 1st Edition, 2019.
3. Matrix Methods: Applied Linear Algebra, Richard Bronson and Gabriel B. Costa, Academic Press; 3rd Edition,
2008.
4. Matrix Methods in Data Mining and Pattern Recognition (Fundamentals of Algorithms), Lars Eldén – Society
for Industrial and Applied Mathematics, 2007.

Total
L T P C
AML 5102: Applied Machine Learning hours

3 0 0 3 36

Course Outcome

1. Differentiate between different types of machine learning paradigms and choose an appropriate one for a
given application problem.
2. Apply different types of supervised and unsupervised machine learning algorithms to practical problems
and assess their performance.
3. Understand the importance of feature engineering in machine learning applications.

Unit No. of
Topics
Hours

Introduction to Machine Learning; Introduction to Supervised Learning; Decision

Trees:
Overview of Supervised (regression and classification), unsupervised (clustering and
dimensionality reduction), semi-supervised, and reinforcement learning with practical

I examples - Machine learning nomenclature: raw data, types of features and outputs,
feature vector. 12

Computing distances and similarities - Prototype based classification - K-nearest

neighbours - Over- and under-fitting -Introduction to cross validation.
Decision tree model of learning - Classification and regression using decision trees -
Splitting criteria: entropy, information gain, Gini impurity - Building a decision tree.
Linear Models; Feature Selection; Introduction to Unsupervised Learning:
Linear model for regression and classification - Simple linear regression: model,
II estimation and interpretation of coefficients - Introduction to bias/variance trade-off - 12
Regularized linear regression.
Filter, wrapper, and embedded methods.

ME (Artificial Intelligence and Machine Learning) 8

 Clustering vs. classification - Hierarchical clustering: dendrogram construction, types
of linkage - Dimension reduction using principal component analysis (PCA)
Probabilistic Models for Supervised Learning; Support Vector Machine; Ensemble
Methods:
Probabilistic modelling of data using parameters - Introduction to maximum likelihood
III estimation (MLE) of parameters - Naive Bayes model for classification - Logistic 12
regression for binary classification
Classification using linear SVM - Dealing with nonlinearly separable data
Bagging: classification using random forest - Boosting
References

1. Grokking Machine Learning, Luis G. Serrano, Manning Publications; 1st Edition, 2019.
2. Online resource from Manning Publications available at https://www.manning.com/books/grokking-
machine-learning
3. A Course in Machine Learning, Hal Daumé III – Online resource available at http://ciml.info/
4. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor Hastie
and Robert Tibshirani, Springer; 1st Edition, 2013, Corr. 7th printing 2017 Edition.
5. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf.

L T P C Total hours
AML 5103: Applied Probability and Statistics
3 0 0 3 36

Course Outcome

1. Model random phenomena using random variables.

2. Construct Bayesian models for quantifying uncertainty in practical problems.
3. Use sample information and perform hypothesis-test analysis using an appropriate statistical technique to
explain attributes of a population.
Unit
Topics No. of Hours

Counting; Probability Concepts; Conditional Probability:

I
Multiplication rule; permutation; combination - Sampling: with/without replacement 8
and order matters/does not matter - Binomial & multinomial coefficients - Distribution

ME (Artificial Intelligence and Machine Learning) 9

 problems; Set theory; sample space; outcomes; events - Frequency based definition of
probability - Equally likely vs. not equally likely outcomes - Axioms of probability
Conditional probability; probability tree model; chain rule - Decomposition and the law
of total probability - Bayes’ rule - intuition, dependence/independence of events.
Random Variables:
Modelling using discrete random variables: Bernoulli, geometric, binomial, negative
binomial, hypergeometric, and Poisson distributions - Probability mass function and
II cumulative distribution function - Expectation and variance: discrete case - Modelling 16
using continuous random variables: uniform, normal, log-normal, exponential, and beta
distributions; probability density function - Expectation and variance: continuous case -
Functions of random variables.
Sampling and Parameter Estimation:
Population and sample - Statistic & sampling distribution - Sample mean and variance -
Central limit theorem – intuition and applications
III 12
Point estimation - Standard error - Interval estimation: interpretation of confidence
interval - Hypothesis testing: p-values, significance level and their interpretations,
application to analysis of one- /two-sample mean and paired data.
References

1. Digital Dice: Computational Solutions to Practical Probability Problems, Paul Nahin, Princeton University
Press
2. Introduction to Probability, Charles M. Grinstead, American Mathematical Society; 2nd Revised Edition
1997. Available online at
https://chance.dartmouth.edu/teaching_aids/books_articles/probability_book/amsbook.mac.pdf
3. A First Course in Probability, Sheldon Ross, 9th Edition, Pearson Education India; 9th Edition, 2013
4. Statistics without Tears: An Introduction for Non-Mathematicians (Paperback), Derek Rowntree, Penguin
UK
5. Biostatistics Open Learning textbook – Online resource from University of Florida available at
https://bolt.mph.ufl.edu/6050-6052/
6. All of Statistics: A Concise Course in Statistical Inference, Larry Wasserman – Springer

ME (Artificial Intelligence and Machine Learning) 10

 ELECTIVES - SEMESTER I

L T P C Total hours
AML 5131: Applications of Graph Theory
3 0 0 3 36

Course Outcome

1. Develop a thorough understanding of fundamental graph theoretic concepts and apply them to
understanding practical problems.
2. Relate a real-life problem to an appropriate graph theoretic setup.
3. Describe how graph theory can be used for machine learning applications.
Unit Topics No. of Hours

Graphs; Euler Tours and Hamilton Cycles:

Graphs and their representations - Incidence and adjacency matrices - Vertex degrees
I 14
- Paths and connection - Cycles - Directed graphs - Subgraphs and supergraphs - The
shortest path problem - Forests and trees, Cayley’s formula.

Flow in Networks; Matchings; Colouring Problems:

Flows and cuts - Max-flow min-cut theorem and its applications.
II 14
Matchings and coverings in bipartite graphs - Perfect matchings - Applications of
matchings. Edge colouring & Vertex colouring

III Random walks and Applications; Spectral Clustering and Applications 8

References

1. Introduction to Graph Theory, Richard J. Trudeau, Dover Publications Inc.: 2nd Revised Edition, 1994.
2. Pearls in Graph Theory: A Comprehensive Introduction, Nora Hartsfield and Gerhard Ringel, Dover
Publications, 2003.
3. Graph Theory, Adrian Bondy, M. Ram Murty, Springer Publications,1st Edition, 2008.

ME (Artificial Intelligence and Machine Learning) 11

 Total
L T P C
hours
BDA 5102: Architecture of Big Data Systems

3 0 0 3 36

Course Outcome

1. Examine different types of data and understand lambda architecture.

2. Understanding different tools and frameworks of Hadoop eco-system.
3. Understanding real-time data processing using Spark engine.
4. Design applications to handle batch and streaming data using Hadoop and Spark tools.

No. of
Unit Topics Hours

Classifying Big Data Characteristics and Big Data processing - the Lambda
architecture:
I Analysis type, Processing methodology, Data Types, Data sources 12
Different data storing and processing layers and architecture: Batch layer, Serving
layer and Speed layer
Batch layer, Serving layer and Speed layer:
Choosing a storage solution for the batch layer: Distributed file systems, Vertical
II 9
partitioning. MapReduce: a paradigm for Big Data computing. Performance metrics
for the serving layer. Speed layer.
Spark: Alternatives to MapReduce:

III Spark Architecture, Spark Session, DataFrame, Transformations and Actions, Spark 9
SQL, Resilient Distributed Datasets (RDDs)
Stream Processing and Machine Learning using Spark:
Advantages and challenges of stream processing, Stream Processing Design Points,
IV Streaming APIs, Structured Stream Processing High Level M-Lib concepts and M-Lib in 6

Action.

References

1. Big Data: Principles and best practices of scalable real-time data systems - Nathan Marz and James Warren.
Manning Publisher.

2. Hadoop: The Definitive Guide: Storage and Analysis at Internet Scale – Tom White, O’Reilly Publication 4th
Edition.

ME (Artificial Intelligence and Machine Learning) 12

3. Spark: The Definitive Guide: Big Data Processing Made Simple – Bill Chambers, Matei Zaharia, O’Reilly
Publication 1st Edition.

L T P C Total hours
BDA 5132: Principles of Data Visualization
3 0 0 3 36

Course Outcome

1. Employ web scrapping techniques to extract data from websites.

2. Illustrate data analysis techniques to prepare data for visualization.
3. Outline different visualization techniques.

No. of
Unit Topics Hours

Introduction to Web scraping:

I Web scraping models and techniques - Case study BeautifulSoup, Scrapy, Selenium. 8

Data Analysis:
Data structures for analysis: numpy, pandas
II 12
Data Wrangling: Clean, Transform, Merge, Reshape, Data Aggregation and Group
Operations, Exploratory analysis of public / scrapped datasets.
Data Visualization:
Data Visualization – classification, infographics versus data visualization, visualization
III for supporting exploratory data analysis, visual art, choosing appropriate visual 16
encodings, Visualization techniques: time series, statistical distributions, maps - Data
visualization for web.

References

1. Website Scraping with Python: Using BeautifulSoup and Scrapy, Gábor & Hajba, APRESS Publications, 1st
Edition, 2018.
2. Web Scraping with Python: Collecting More Data from the Modern Web, Ryan Mitchell Shroff, O'Reilly, 2nd
Edition, 2018.
3. Designing Data Visualizations, Julie Steele and Noah Iliinsky; O'Reilly Media; 1st Edition, 2011.
4. Python for Data Analysis, Wes McKinney; Shroff; O'Reilly; 2nd Edition, 2018.

ME (Artificial Intelligence and Machine Learning) 13

 L T P C Total hours
BDA 5151: Algorithms and Data structures for Big Data
Lab
0 0 3 1 36

Course Outcome

1. Design programs for implementation of linked lists, stack and queues.

2. Design programs for implementation of binary search tree, sorting and searching, dictionary and Hash Table
3. Design programs for graphs and shortest path techniques.

No. of
Unit Topics Hours

Implement single linked and double linked list

I 15
Implement array based and linked list-based Stack
Implement array based and linked list-based Queue

II Implement binary tree and BST

6
Design and Implement Hash functions, implement open and closed Hash Tables

III Implement Graph algorithms and shortest path 15

References

1. Introduction to Algorithms - Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest. MIT Press.
2. Data Structures and Algorithms - Aho, Hopcroft and Ulmann. Pearson Publishers.
3. Data Structures and Algorithms in Python - Michael T. Goodrich, Roberto Tamassia, and Michael H.
Goldwasser. John Wiley & Sons.
4. Data Streams: Algorithms and Applications - S. Muthukrishnan. Foundations and Trends in Theoretical
Computer Science archive, Volume 1 Issue 2, August 2005, Pages 117 – 236

ME (Artificial Intelligence and Machine Learning) 14

 Total
L T P C
hours
AML 5151: Applied Linear Algebra Lab

0 0 3 1 36

Course Outcome

1. Develop solid skills in using Python’s legacy libraries for coding matrix-vector operations.
2. Implement algorithms constructed using matrix-vector principles.

3. Implement models for real-life applications using the least squares technique and interpret the results from
a practical perspective.
No. of
Unit Topics Hours

Vectors:
Understand how to perform vector operations using Python; Visualize vectors and
I relate them to their geometric description; Implement the K-means algorithm from 10
scratch using vector operations; Implement and interpret the output of the Gram-
Schmidt algorithm.
Matrices:
Understand how to perform matrix operations using Python; Implement and interpret
II matrix-vector operations using block-matrix operations; Understand how to solve 14
linear systems of equations using Python; Code practical applications of QR
factorization of matrices.
Linear Least Squares:
Solve linear least squares problems using Python and interpret the results; Implement
III and fine-tune feature extraction using least squares for practical problems; 12

Implement least squares classification.

References

1. Introduction to Applied Linear Algebra, Vectors, Matrices, and Least Squares, Stephen Boyd & Lieven
Vandenberghe, Cambridge University Press, 1st Edition, 2018. Available online at http://vmls-
book.stanford.edu/vmls.pdf

2. Linear Algebra and Learning from Data, Gilbert Strang, Cambridge Uni. Press; 1st Edition, 2019.

3. Matrix Methods: Applied Linear Algebra, Richard Bronson and Gabriel B. Costa, Academic Press; 3rd Edition,
2008.

ME (Artificial Intelligence and Machine Learning) 15

4. Matrix Methods in Data Mining and Pattern Recognition (Fundamentals of Algorithms), Lars Eldén – Society
for Industrial and Applied Mathematics, 2007

L T P C Total hours
AML 5152: Applied Machine Learning Lab
0 0 3 1 36

Course Outcome

1. Differentiate between different types of machine learning paradigms and choose an appropriate one for a
given application problem.
2. Apply different types of supervised and unsupervised machine learning algorithms to practical problems
and assess their performance.
3. Understand the importance of feature engineering in machine learning applications.

Unit No. of
Topics
Hours

Introduction to Machine Learning; Introduction to Supervised Learning; Decision

Trees:
Program data, perform data wrangling, understand the data matrix, and differentiate
I between sample and feature; Investigate over- and underfitting concepts using the K- 12
nearest neighbour algorithm; Implement and interpret results of cross-validation;
Implement decision tree models in Python, fine-tune model parameters, and interpret
results.
Linear Models; Feature Selection; Introduction to Unsupervised Learning:
Implement linear models in Python and interpret model coefficients for practical

II problems; Implement and visualize bias-variance trade-off using linear regression as

12
a basis; Compare, and contrast different feature engineering approaches for practical
problems; Visualize the output of hierarchical clustering and PCA algorithms and
interpret the results.
Probabilistic Models for Supervised Learning; Support Vector Machine; Ensemble
Methods:
III 12
Implement maximum likelihood estimation for a simple model; Analyse the
performance of the Naive Bayes model for practical problems; Apply the SVM

ME (Artificial Intelligence and Machine Learning) 16

 algorithm for linearly- and not-linearly separable data, compare the performance;
Through coding, understand how ensemble methods in machine learning work.

References

1. Grokking Machine Learning, Luis G. Serrano, Manning Publications; 1st Edition, 2019.
2. Online resource from Manning Publications available at https://www.manning.com/books/grokking-
machine-learning
3. A Course in Machine Learning, Hal Daumé III – Online resource available at http://ciml.info/
4. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor Hastie
and Robert Tibshirani, Springer; 1st Edition, 2013, Corr. 7th printing 2017 Edition.
5. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf.

AML 5153: Applied Probability and L T P C Total hours

Statistics Lab
0 0 3 1 36

Course Outcome

1. Visualize probability concepts through frequency-based interpretations.

2. Simulate discrete and continuous random variables for modelling random phenomena.
3. Design and apply hypothesis tests followed by interpretation of results.
Unit
Topics No. of Hours

Counting; Probability Concepts; Conditional Probability:

Understand the basic principles of the R programming language; Develop short code
I
snippets to understand the basic principles of sampling and probability; Visualise and 8
interpret probability concepts through a frequency-based approach; Program and
analyse Bayesian models for practical problems.
Random Variables:
Understand and apply R functions to simulate discrete and continuous random

II variables; Using sampling, compute and interpret different attributes of random

16
variables; Visualise and interpret histograms and probability mass/density functions
of random variables using state of the art visualisation libraries in R; Develop codes
to model random phenomena using appropriate random variables.

ME (Artificial Intelligence and Machine Learning) 17

 Sampling and Parameter Estimation:
Visualise sample data through histograms; Compute estimates of population
III parameters using samples and communicate the uncertainty in the estimates; Use R 12
in-built functions for performing hypothesis tests; Interpret and communicate the
results of hypothesis tests.
References

1. Digital Dice: Computational Solutions to Practical Probability Problems, Paul Nahin, Princeton University
Press
2. Introduction to Probability, Charles M. Grinstead, American Mathematical Society; 2nd Revised Edition
1997. Available online at
https://chance.dartmouth.edu/teaching_aids/books_articles/probability_book/amsbook.mac.pdf
3. A First Course in Probability, Sheldon Ross, 9th Edition, Pearson Education India; 9th Edition, 2013
4. Statistics without Tears: An Introduction for Non-Mathematicians (Paperback), Derek Rowntree, Penguin UK
5. Biostatistics Open Learning textbook – Online resource from University of Florida available at
https://bolt.mph.ufl.edu/6050-6052/
6. All of Statistics: A Concise Course in Statistical Inference, Larry Wasserman – Springer

L T P C Total hours
AML 5181: Applications of Graph Theory Lab
0 0 3 1 36

Course Outcome

1. Visualize graphs and graph models using Python.

2. Implement appropriate algorithms for solving graph theoretical problems.
3. Implement graph theoretic approaches for machine learning applications.

Unit Topics No. of Hours

Graphs; Euler Tours and Hamilton Cycles:

I Visualize graphs using Python; Compute structural properties of graphs such as 14

connectivity using the adjacency matrix.

Flow in Networks; Matchings; Colouring Problems:

II Visualize network flow problems through applications; Implement practical 14
applications of matching, edge- & vertex-colouring.

ME (Artificial Intelligence and Machine Learning) 18

 Random walks and Applications; Spectral Clustering and Applications:
III Create a random walk and analyse its properties; Model multidimensional data as 8
similarity graph; Apply spectral clustering to practical problems.

References

1. Introduction to Graph Theory, Richard J. Trudeau, Dover Publications Inc.: 2nd Revised Edition, 1994.
2. Pearls in Graph Theory: A Comprehensive Introduction, Nora Hartsfield and Gerhard Ringel, Dover
Publications, 2003.
3. Graph Theory, Adrian Bondy, M. Ram Murty, Springer Publications,1st Edition, 2008.

L T P C Total hours
BDA 5152: Architecture of Big Data Systems Lab
0 0 3 1 36

Course Outcome

1. Identify and work with different types of data.

2. Work with different tools and frameworks of Hadoop eco-system.
3. Handle real-time data processing using Spark engine.
4. Design applications to handle batch and streaming data using Hadoop and Spark tools.

No. of
Unit Topics Hours

Using HDFS commands to get familiarize with HDFS environment

I Transfer structured data to and from HDFS

15
Creating Hive tables in HDFS and running SQL statements
Developing and running MapReduce programs

Creating RDDs in spark

II Creating Spark DataFrames, loading, transforming and performing actions in Spark

12
environment
Handling streaming data using structured streaming

Developing applications using Hadoop tools

III 9
Develop applications using Spark environment

References

1. Big Data: Principles and best practices of scalable real-time data systems - Nathan Marz and James Warren.
Manning Publisher.

ME (Artificial Intelligence and Machine Learning) 19

2. Hadoop: The Definitive Guide: Storage and Analysis at Internet Scale – Tom White, O’Reilly Publication 4th
Edition.
3. Spark: The Definitive Guide: Big Data Processing Made Simple – Bill Chambers, Matei Zaharia, O’Reilly
Publication 1st Edition.

L T P C Total hours
BDA 5182: Principles of Data Visualization Lab
0 0 3 1 36

Course Outcome

1. Experiment web scrapping techniques to extract data from websites.

2. Prepare data for visualization using data analysis techniques.
3. Model different visualization techniques.

No. of
Unit Topics Hours

I Design programs to dynamically extract data from web. 12

Understand and integrate various data structures for data analysis process.
Create various techniques to clean and handle missing data.
II 9
Design data filtering and transformation techniques.

Describe what is the purpose of Visualization and ways of classifying visualization.

III Create visualization for time series data, statistical distributions. 15

Create visualization for maps, Hierarchical data and network data.

References

1. Website Scraping with Python: Using BeautifulSoup and Scrapy, Gábor & Hajba, APRESS Publications, 1st
Edition, 2018.
2. Web Scraping with Python: Collecting More Data from the Modern Web, Ryan Mitchell Shroff, O'Reilly,
2nd Edition, 2018.
3. Designing Data Visualizations, Julie Steele and Noah Iliinsky; O'Reilly Media; 1st Edition, 2011.
4. Python for Data Analysis, Wes McKinney; Shroff; O'Reilly; 2nd Edition, 2018.

ME (Artificial Intelligence and Machine Learning) 20

 L T P C Total hours
MPT 5100: Mini Project - I
0 0 0 4 48

Course Outcome

1. Identify the real-world and social relevant problems and perform feasibility analysis for finding
solution.
2. Develop solutions to the identified problems by applying research methodology and development life
cycle with appropriate documentation by incorporating ethical standards.
3. Work effectively as a member in a team and communicate technical information effectively.

No. of
Unit Topics
Hours

I Problem identification, literature survey, formation of detailed specifications.

II Design and implementation of the proposed system architecture. 48

III Demonstrate an ability to present and defend project work carried out to a panel
of experts.

References

1. Research articles and Online Resources.

L T P C Total hours
PSD 5100: Professional Skill Development - I
0 0 0 1 12

Course Outcome

1. Identify and synthesize important themes in the field of engineering which transform socio-economic
ecosystem.
2. Develop competence to communicate effectively in oral and written forms.

ME (Artificial Intelligence and Machine Learning) 21

3. Effective management of time, involve in reflective learning and adhere to the professional code of
conduct.

No. of
Unit Topics
Hours

Report writing involves identifying the topic of interest from current issues in the
domain of engineering and technology or inter disciplinary domains, then framing
I
the order in the report, writing abstract, deciding on the content itself, conclusion
and future scope of the topic and properly citing the references from bibliography. 12
Presenting in classroom to audience where content spoken, the conceptual
II knowledge and presentation skills (like audibility, eye contact, memory) of speaker
is assessed.

References

1. Research articles and Online Resources.

ME (Artificial Intelligence and Machine Learning) 22

 SEMESTER II

AML 5201: Advanced Applications of Probability & L T P C Total hours

Statistics
3 0 0 3 36

Course Outcome

1. Apply linear and logistic regression models for practical problems and assess model performance.
2. Interpret the output of principal component analysis (PCA) applied to multivariate data for dimension
reduction.
3. Identify multivariate data with mixed data type features and cluster using an appropriate technique.
4. Understand the basics of time series modelling and apply to real-life problems.
Unit
Topics No. of Hours

Multivariate Distributions:
Mean vector, covariance and correlation – population vs. sample - The

I multivariate Gaussian – joint-, marginal-, and conditional distributions,

Mahalanobis distance and outliers - Properties of the multivariate Gaussian - 6

Parameter estimation: maximum likelihood estimation (MLE) and maximum

aposteriori estimation (MAP).
Linear and Logistic Regression:
Simple linear regression – regression model, estimating and interpreting
coefficients, accuracy of coefficient estimates and model, ANOVA, R2 statistic

II - Multiple linear regression – estimating coefficients, qualitative predictors,

12
interaction effects, potential problems - Logistic regression – binary and
multinomial logistic regression models, estimating and interpreting
coefficients, assessing model calibration and discrimination, area under the
ROC.
Principal Component Analysis; Cluster Analysis:
Geometric intuition of principal components - Maximum variance perspective
– algebraic setup, eigenvectors and eigenvalues of sample correlation matrix -
III Interpretation and application of principal components for dimension 10
reduction.
Dissimilarity measures for mixed data types - Partition around medoids (PAM)
vs. K-means algorithms - Selecting the number of clusters.

ME (Artificial Intelligence and Machine Learning) 23

 Bootstrapping; Time Series Analysis:
Time series concepts: stationarity, trend, seasonality, autocorrelation -
IV Autoregressive moving average (ARMA) models - Resampling, smoothing, 8
windowing, and rolling average - First and second order differencing -
Validating time series predictions.
References

1. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor Hastie
and Robert Tibshirani, Springer; 1st Edition, 2013, Corr. 7th printing 2017 Edition.

2. An Introduction to Applied Multivariate Analysis with R, Brian Everitt and Torsten Hothorn– Springer
Publications,1st Edition, 2011.

3. Machine Learning - A Probabilistic Perspective, Kevin P. Murphy, The MIT Press; 1st Edition, 2012.

4. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong, Cambridge
University Press, 2020. – Online resource from Cambridge University Press available at https://mml-
book.github.io/book/mml-book.pdf

L T P C Total hours
AML 5202: Deep Learning
3 0 0 3 36

Course Outcome

1. Gain a solid understanding of the mathematical basis of neural networks.

2. Build and analyse deep learning models for application problems.
3. Devise techniques for improving the way neural networks learn.
4. Develop skills to choose an appropriate deep learning model.
Unit
Topics No. of Hours

Introduction to Deep Learning; Matrix Calculus; Logistic Regression:

Sigmoid neurons - The architecture of neural networks.
I Derivatives in one dimension - Derivative in multiple dimensions: gradient and 9
Jacobian matrices - Rules of matrix calculus: product and chain rules -
Optimizing using the gradient descent method– intuition and principle.

ME (Artificial Intelligence and Machine Learning) 24

 Binary classification using logistic regression: cost function, gradient descent,
and vectorization.
Shallow Neural Network:
One hidden layer neural network: architecture and notation - The role of

II activation functions and their derivatives - Forward propagation using matrix-

10
based approach - Cost/loss function: intuition and setup - Gradient descent:
backpropagation intuition and vectorized setup using matrix-based approach
- Random initialization of network parameters.
Deep Neural Network; Improving the Way neural Networks Learn:
Deep L-layer neural network: architecture, notation, and building blocks -
Forward and backward propagation in a deep neural network using matrix-
III based approach - The importance of deep representations - Parameters vs. 11
hyperparameters.
The cross-entropy cost function - The learning slowdown problem - Overfitting
and regularization: L1/L2, dropout - Weight initialization.
Hyperparameter Tuning; Recurrent Neural Networks:
Random initialization using appropriate scales - Batch normalization.
IV Recurrent neural network: architecture and notation - Forward and backward 6
propagation through time - Different types of recurrent neural networks and
their applications.
References

1. Neural Networks and Deep Learning, Michael Nielsen – Determination Press – Available online at
http://neuralnetworksanddeeplearning.com/index.html

2. Lecture slides of Prof. Andrew Ng – Stanford University – Available online at

https://cs230.stanford.edu/syllabus/

3. Tensor Flow for Deep Learning Paperback, Reza Zadeh & Bharath Ramsundar, O'Reilly, 2018.

4. Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville – MIT Press – Available online at
http://www.deeplearningbook.org/

ME (Artificial Intelligence and Machine Learning) 25

 L T P C Total hours
AML 5203: Machine Learning Principles & Applications
3 0 0 3 36

Course Outcome

1. Differentiate between discriminative and generative algorithms for supervised machine learning.
2. Evaluate machine learning algorithms for accuracy and performance.
3. Devise techniques for dealing with practical difficulties in applying machine learning techniques to real-
life problems.
4. Develop low dimensional models of application problems with mixed data type features.
Unit Topics No. of Hours

Kernel Methods; Linear Regression:

Kernels as feature maps - Kernel functions: types, hyperparameters - Kernel

I matrix: interpretation and properties - Kernel (nonlinear) SVM.

Least mean squares (LMS) algorithm: cost function - Gradient descent 10

algorithm: learning rate, batch and stochastic gradient approaches -

Probabilistic interpretation of linear regression: MLE and MAP estimates.
Generative Learning Algorithms; Regularization, Model Selection, &
Evaluation:
Gaussian discriminant analysis (GDA) - Naive Bayes algorithm: MLE estimates,

II Laplace smoothing.
11
Grid search for best hyperparameters - Cross validation: types and practical
approaches – Feature selection: forward/backward search, wrapper model &
filter feature selection - Metrics for evaluating supervised & unsupervised
machine learning algorithms.
Imbalanced Data; Expectation Maximization; Dimension Reduction;
Independent Component Analysis:
Modifying the training data: over- and under-sampling - Modifying the loss
function.
III Clustering with a mixture of Gaussians - Expectation maximization (EM) 15
framework.
Factor analysis (FA) - Generalized low rank models (GLRM).
Independent Component Analysis (ICA)

ME (Artificial Intelligence and Machine Learning) 26

References

1. A Course in Machine Learning, Hal Daumé III – Online resource available at http://ciml.info/

2. Machine Learning: A Probabilistic Perspective, Kevin Murphy, MIT Press,2017.

3. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor Hastie
and Robert Tibshirani, Springer Publications, 2017.

4. Lecture slides of Prof. Andrew Ng – Stanford University – Available online at

http://cs229.stanford.edu/syllabus.html

5. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf

6. Pattern Recognition and Machine Learning, Christopher Bishop, Springer Publications, 2017.

L T P C Total hours
AML 5204: Reinforcement Learning
3 0 0 3 36

Course Outcome

1. Define the key features of reinforcement learning that distinguishes it from AI and non-interactive machine
learning.
2. Understand how ideas such as temporal difference learning and dynamic programming fit in the framework
of learning from interaction to achieve goals.
3. Decide if an application problem can be formulated as a reinforcement learning problem and choose an
appropriate algorithm.
4. Understand and implement commonly used reinforcement learning algorithms.
Unit
Topics No. of Hours

Introduction to the Reinforcement Learning Problem; Reinforcement

I Learning Framework; Dynamic Programming:

Examples and elements of reinforcement learning - Limitations and scope of 15

reinforcement learning - History of reinforcement learning.

ME (Artificial Intelligence and Machine Learning) 27

 n-Armed bandit problem: action-value methods - Finite Markov decision
process: the agent–environment interface, goals and rewards, returns, Markov
decision processes, value functions, and optimal value functions.
Model Free Reinforcement Learning:
II 11
Generalized policy iteration - Importance of exploration - Monte Carlo control
- Temporal difference methods for control.
Approximate Solution Methods; Policy Based Methods:
Value prediction with function approximation - Gradient-descent methods -
III Linear methods - Control with function approximation. 10
Policy gradient - Actor–critic methods - Policy-based vs. value-based methods
- Integrating supervised & reinforcement learning.
References

1. Deep Reinforcement Learning Hands-On, Maxim Lapan, 2nd Edition, Packt, 2020.

2. Reinforcement Learning: An Introduction, Richard S. Sutton and Andrew G. Barto, MIT Press, 2nd Edition –
Available online at https://web.stanford.edu/class/psych209/Readings/SuttonBartoIPRLBook2ndEd.pdf

3. Generative Deep Learning: Teaching Machines to Paint, Write, Compose, and Play, David Foster – O’Reilly,
1st Edition, 2019.

ME (Artificial Intelligence and Machine Learning) 28

 ELECTIVES - SEMESTER II

Total
L T P C
hours
AML 5231: Applied Mathematics for Machine Learning

3 0 0 3 36

Course Outcome

1. Develop a solid understanding of fundamentals of matrix decomposition techniques and apply them to
practical problems.
2. Describe the role of derivatives in machine learning and understand different methods for computing
them.
3. Acquire solid foundation in understanding the principles behind state-of-the-art optimization algorithms
used in machine learning libraries.

Unit No. of
Topics
Hours

Matrix Decompositions and Applications:

I Matrix and tensor products - Determinant and trace - Eigen-decomposition and

diagonalization - Cholesky decomposition - Singular value decomposition - 14

Nonnegative matrix factorization.

Computing Derivatives:
II 8
Differentiability - Symbolic differentiation - Finite differences - Automatic
differentiation.

Continuous Optimization:
Optimization using gradient descent - Constrained optimization and Lagrange
III 14
multipliers - Convex optimization – Sub gradients - Stochastic gradient descent
- Momentum methods.

ME (Artificial Intelligence and Machine Learning) 29

References

1. Mathematics for Machine Learning by Marc Peter Deisenroth, Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf

2. Matrix Computations, Gene H. Golub and Charles F. Van Loan, Hindustan Book Agency; 4th Edition, 2015.

3. Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville, MIT Press,2017. – Available online
at http://www.deeplearningbook.org/

4. Understanding Machine Learning: From Theory to Algorithms (UML), Shai Shalev-Shwartz and Shai Ben-
David, Cambridge University Press, 1st Edition, 2014.

Total
L T P C
AML 5232: Convolutional Neural Networks for Computer hours
Vision
3 0 0 3 36

Course Outcome

1. Analyze a real-life problem involving computer vision and solve it using CNNs.
2. Develop and design CNN architectures using the basic building blocks of CNN
3. Decide how to choose an existing CNN architecture for an application problem.
Unit
Topics No. of Hours

Introduction to Computer Vision; Features; Neural Networks Basics:

Computer vision overview - Historical context and applications - Image
I
processing vs. computer vision 6
Role of convolution in image processing, neighbourhood operations for image
enhancement, edge detection, colour models – RGB, HSI, CIE LAB, YUV.
Convolutional Neural Networks (CNN):

II Network layers: pre-processing, convolutional layers, pooling layers,

10
nonlinearity, fully connected layers, region of interest pooling - Loss functions:
hinge loss, squared hinge loss, cross-entropy loss, Euclidean loss, L1 error.
CNN Learning; Visualizing and Understanding CNNs:
III Weight initialization – Regularization - Gradient based learning: batch-, 10
stochastic-, and mini-batch gradient descent, gradient computations in CNN.

ME (Artificial Intelligence and Machine Learning) 30

 CNN Architectures; Applications of CNNs in Computer Vision:
IV 10
Image classification, Object detection and localization.
References

1. A Guide to Convolutional Neural Networks for Computer Vision, Salman Khan, Hossein Rahmani, Syed Afaq
Ali Shah, and Mohammed Bennamoun, Morgan & Claypool Publishers, 2018.

2. Lecture slides of Prof. Fei-Fei Li – Stanford University – Available online at http://cs231n.stanford.edu/

3. Neural Networks and Deep Learning, Michael Nielsen, Determination Press – Available online at
http://neuralnetworksanddeeplearning.com/index.html

4. Computer Vision: Algorithms and Applications, Richard Szeliski, Springer, 2011 – Online resource from
Springer available at http://szeliski.org/Book/

5. Lecture slides of Prof. Fei-Fei Li – Stanford University – Available online at http://cs231n.stanford.edu/

L T P C Total hours
AML 5233: Natural Language Processing Principles &
Applications
3 0 0 3 36

Course Outcome

1. Develop an in-depth understanding of both algorithms for processing linguistic information and the
underlying computational properties of natural languages.
2. Analyse word-level, syntactic, and semantic processing from both a linguistic and an algorithmic
perspective.
3. Formulate deep learning approaches for natural language processing tasks.
Unit
Topics No. of Hours

Introduction to Natural Language Processing (NLP); Regular Expressions; N-

gram Language Models:
I
Terminology - Probability and NLP. 14
Introduction to regular expressions - Information extraction using regular
expressions.

ME (Artificial Intelligence and Machine Learning) 31

 Probabilistic language model - Chain rule and Markov assumption - Evaluating
language models – Smoothing.
Naive Bayes and Sentiment Classification; Vector Semantics and
Embeddings:
II 8
Vector semantics - Words and vectors - Cosine for measuring similarity - TF-IDF
vector model - Word2Vec &
GloVe models - Visualizing embeddings.
NLP with Deep Learning; Applications of Natural Language Processing:
III Neural language models - Introduction to PyTorch -Sequence processing with 14
recurrent neural networks.
References

1. Speech and Language Processing, Dan Jurafsky and James H. Martin, Pearson; 3rd Edition (draft) – Available
online at https://web.stanford.edu/~jurafsky/slp3/

2. Natural Language Processing with Python. – Analysing Text with the Natural Language Toolkit, Steven Bird,
Ewan Klein, and Edward Loper, ISTE Ltd., 1st Edition, 2017 - Available online at
https://www.nltk.org/book/

3. A Primer on Neural Network Models for Natural Language Processing, Yoav Goldberg – Available online at
http://faculty.cse.tamu.edu/huangrh/Spring18/nnlp.pdf

4. Natural Language Processing with PyTorch, Delip Rao & Brian McMahan, O'Reilly, 1st Edition, 2019.

L T P C Total hours
ENP 5230: Entrepreneurship
3 0 0 3 36

Course Outcome
1. Explain the importance of entrepreneurship and entrepreneurial development model, social
responsibilities of business.
2. Describe Entrepreneurial Traits and Factors affecting Entrepreneurship process.
3. Discuss Business Start-up Process.
4. Summarize a business and marketing plan for entrepreneurs.

ME (Artificial Intelligence and Machine Learning) 32

Unit Topics No. of Hours

Introduction to Entrepreneurship:
Meaning and Definition of Entrepreneurship-Employment vs Entrepreneurship,
I 6
Theories of Entrepreneurship, approach to entrepreneurship, Entrepreneurs VS
Manager

Entrepreneurial Traits:
II 5
Personality of an entrepreneur, Types of Entrepreneurs

Process of Entrepreneurship:
III 6
Factors affecting Entrepreneurship process
Business Start-up Process:
IV 7
Idea Generation, Scanning the Environment, Macro and Micro analysis
Business Plan writing:
V 6
Points to be considered, Model Business plan
Case studies:
VI 6
Indian and International Entrepreneurship
References
1. NVR Naidu and T. Krishna Rao, “Management and Entrepreneurship”, IK International Publishing House
Pvt. Ltd 2008.
2. Mohanthy Sangram Keshari, “Fundamentals of Entrepreneurship”, PHI Publications, 2005

AML 5251: Advanced Applications of L T P C Total hours

Probability & Statistics Lab
0 0 3 1 36

Course Outcome

1. Build and assess linear and logistic regression models for practical problems.
2. Perform principal component analysis (PCA) for dimension reduction in multivariate data.
3. Cluster multivariate data with mixed data types.
4. Apply time series modelling to real-life problems.

ME (Artificial Intelligence and Machine Learning) 33

Unit No. of
Topics
Hours

Multivariate Distributions:

I Compute descriptive statistics of multivariate data; Perform exploratory data analysis

of multivariate data; Identify outliers in multivariate data; Visualise and understand 6

the properties of multivariate Gaussian data.

Linear and Logistic Regression:
Use in-built functions in R to build linear models for practical problem; Compute

II different performance metrics to assess model performance; Interpret model

12
coefficients and investigate the effect of input variables on output through sensitivity
analysis; Use in-built functions in R to build logistic regression models for practical
binary classification problems and assess model performance.
Principal Component Analysis; Cluster Analysis:
Visualise the geometric interpretation of principal component analysis (PCA); Use in-
III built functions in R to perform PCA on multivariate data; Compare and contrast PCA 10
for variance maximization vs. clustering of multivariate data; Cluster multivariate
data with mixed data types using in-built functions in R.
Bootstrapping; Time Series Analysis:
Apply bootstrapping on a practical data set and assess performance; Understand and
IV 8
apply in-built functions in R for time series modelling; Apply time series modelling to
practical problems; Interpret the results of times series model predictions.
References

1. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor
Hastie and Robert Tibshirani, Springer; 1st Edition, 2013, Corr. 7th printing 2017 Edition.

2. An Introduction to Applied Multivariate Analysis with R, Brian Everitt and Torsten Hothorn– Springer
Publications,1st Edition, 2011.

3. Machine Learning - A Probabilistic Perspective, Kevin P. Murphy, The MIT Press; 1st Edition, 2012.

4. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong, Cambridge
University Press, 2020. – Online resource from Cambridge University Press available at https://mml-
book.github.io/book/mml-book.pdf

ME (Artificial Intelligence and Machine Learning) 34

 L T P C Total hours
AML 5252: Deep Learning Lab
0 0 3 1 36

Course Outcome

1. Develop practical experience with state-of-the-art deep learning tools and libraries.
2. Implement deep learning models for application problems.
3. Implement techniques for improving the way neural networks learn.
4. Computationally analyse deep learning models and select the best model.

Unit No. of
Topics
Hours

Introduction to Deep Learning; Matrix Calculus; Logistic Regression:

Implement a sigmoid neuron from scratch; Implement forward and backward
I propagation for a sigmoid neuron; Implement gradient descent for a sigmoid neuron; 9
Implement cost function for binary classification using logistic regression using
vectorized approach.
Shallow Neural Network:
Visualize different nonlinear activation functions; Implement forward and backward
II propagation for a shallow neural network using matrix-based approach; Implement 10
gradient descent method for a shallow neural network; Numerically investigate the
effect of random initialization of network parameters.
Deep Neural Network; Improving the Way neural Networks Learn:
Visualise architecture of a deep neural network; Implement forward and backward
III propagation for a deep neural network using matrix-based approach; Implement deep 11
neural networks using in-built libraries for real-life problems; Implement different
regularization approaches and compare their advantages and disadvantages.
Hyperparameter Tuning; Recurrent Neural Networks:
Fine tune hyperparameters; Numerically investigate the effect of random
IV 6
initialization in deep neural networks; Visualise the architecture of a recurrent neural
network; Implement recurrent neural network models for real-life problems.
References

1. Neural Networks and Deep Learning, Michael Nielsen – Determination Press – Available online at
http://neuralnetworksanddeeplearning.com/index.html

ME (Artificial Intelligence and Machine Learning) 35

2. Lecture slides of Prof. Andrew Ng – Stanford University – Available online at
https://cs230.stanford.edu/syllabus/

3. Tensor Flow for Deep Learning Paperback, Reza Zadeh & Bharath Ramsundar, O'Reilly, 2018.

4. Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville – MIT Press – Available online at
http://www.deeplearningbook.org/

Total
L T P C
AML 5253: Machine Learning Principles & Applications hours
Lab
0 0 3 1 36

Course Outcome

1. Practically apply state of the art machine learning tools and libraries.
2. Evaluate machine learning algorithms for accuracy and performance for practical problems.
3. Implement different strategies for selecting features and dealing with missing data.
4. Implement machine learning models for real-life data with mixed datatype features.

Unit No. of
Topics
Hours

Kernel Methods; Linear Regression:

Implement and compare different kernels for feature mapping; Implement
kernel SVM and investigate the effects of model parameters through
I visualization; Implement gradient descent for least mean squares algorithm 10
and investigate the effects of hyperparameters on performance; Compare
linear regression applied to practical problems with and without
regularization.
Generative Learning Algorithms; Regularization, Model Selection, &
Evaluation:

II Implement probabilistic models of data; Perform grid search to identify best

11
model hyperparameters; Perform feature engineering for real-life problems;
Evaluate machine learning algorithms using well established performance
metrics.
Imbalanced Data; Expectation Maximization; Dimension Reduction;
III 15
Independent Component Analysis:

ME (Artificial Intelligence and Machine Learning) 36

 Implement and compare different approaches for dealing with missing data in
real-life problems; Implement and interpret low dimensional models of data;
Implement models for analysing mixed datatype data; Compare and contrast
different techniques for dimension reduction and their practical implications.
References

1. A Course in Machine Learning, Hal Daumé III – Online resource available at http://ciml.info/

2. Machine Learning: A Probabilistic Perspective, Kevin Murphy, MIT Press,2017.

3. An Introduction to Statistical Learning with Applications in R, Gareth James, Daniela Witten, Trevor
Hastie and Robert Tibshirani, Springer Publications, 2017.

4. Lecture slides of Prof. Andrew Ng – Stanford University – Available online at

http://cs229.stanford.edu/syllabus.html

5. Mathematics for Machine Learning, Marc Peter Deisenroth, A Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf

6. Pattern Recognition and Machine Learning, Christopher Bishop, Springer Publications, 2017.

L T P C Total hours
AML 5254: Reinforcement Learning Lab
0 0 3 1 36

Course Outcome

1. Understand the trade-off between exploration vs. exploitation approaches in solving reinforcement
learning tasks.
2. Use dynamic programming approach to solve reinforcement learning tasks.
3. Model real-life problems using Markov decision processes.
4. Compare and contrast several methods for solving reinforcement learning tasks.

Unit No. of
Topics
Hours

I Introduction to the Reinforcement Learning Problem; Reinforcement Learning

Framework; Dynamic Programming: 15

ME (Artificial Intelligence and Machine Learning) 37

 Implement building blocks for solving a reinforcement learning task; Solve an n-
Armed bandit problem using different exploration strategies; Implement Markov
decision process models.
Model Free Reinforcement Learning:

II Implement iterative policy evaluation; Implement Monte Carlo methods for solving
11
reinforcement learning tasks; Implement temporal difference methods for solving
reinforcement learning tasks.
Approximate Solution Methods; Policy Based Methods:
Implement function approximation methods for value prediction; Implement
III 10
linear value function approximators; Explore integration of supervised and
reinforcement learning.
References

1. Deep Reinforcement Learning Hands-On, Maxim Lapan, 2nd Edition, Packt, 2020.

2. Reinforcement Learning: An Introduction, Richard S. Sutton and Andrew G. Barto, MIT Press, 2nd Edition –
Available online at https://web.stanford.edu/class/psych209/Readings/SuttonBartoIPRLBook2ndEd.pdf

3. Generative Deep Learning: Teaching Machines to Paint, Write, Compose, and Play, David Foster – O’Reilly,
1st Edition, 2019.

Total
L T P C
AML 5281: Applied Mathematics for Machine Learning hours
Lab
0 0 3 1 36

Course Outcome

1. Assess applicability of matrix decomposition techniques for practical problems.

2. Implement and compare different methods for computing derivatives.
3. Understand the implementations of state-of-the-art optimization algorithms used in machine learning
libraries.

ME (Artificial Intelligence and Machine Learning) 38

 Unit No. of
Topics
Hours

Matrix Decompositions and Applications:

Implement matrix decompositions using block matrix representations; Implement
I
and compare exact and approximate decompositions; Implement codes to 14
understand the optimization-centric view to matrix factorization; Interpret the
factors arising out of matrix factorizations for real-life problems.
Computing Derivatives:
Visualize differentiability concepts in 3D; Implement symbolic differentiation for
II 8
computing derivatives exactly; Implement finite difference methods for
approximating derivatives; Implement automatic differentiation and compare it
with other approaches.
Continuous Optimization:
Solve continuous optimization problems using state of the art libraries; Visualize
constrained optimization problems and solutions in 3D; Implement and visualize
III 14
solutions of gradient descent method and its extensions for continuous
optimization; Understand implementations of continuous optimization algorithms
used in state-of-the-art libraries.
References

1. Mathematics for Machine Learning by Marc Peter Deisenroth, Aldo Faisal, and Cheng Soon Ong – Online
resource from Cambridge University Press available at https://mml-book.github.io/book/mml-book.pdf

2. Matrix Computations, Gene H. Golub and Charles F. Van Loan, Hindustan Book Agency; 4th Edition, 2015.

3. Deep Learning, Ian Goodfellow, Yoshua Bengio, and Aaron Courville, MIT Press,2017. – Available online at
http://www.deeplearningbook.org/

4. Understanding Machine Learning: From Theory to Algorithms (UML), Shai Shalev-Shwartz and Shai Ben-
David, Cambridge University Press, 1st Edition, 2014.

ME (Artificial Intelligence and Machine Learning) 39

 L T P C Total hours
AML 5282: Convolutional Neural Networks for Computer
Vision Lab
0 0 3 1 36

Course Outcome

1. Design computer vision techniques for problem solving.

2. Develop and design CNN architectures using the basic building blocks of CNN.

3. Design and develop a CNN architecture for an application problem.

Unit No. of
Topics
Hours

Introduction to Computer Vision; Features; Neural Networks Basics:

I Implement basic computer vision and image processing tasks; Implement feature
extraction using convolution from images; Implement and understand various 6

colour model representations.

Convolutional Neural Networks (CNN):
Visualize and understand the building blocks of a CNN; Implement different types
II 10
of CNN layers and understand their utility. Implement different nonlinear
activation functions, compare and contrast them; Implement and understand the
role of different types of loss functions used in a CNN.
CNN Learning; Visualizing and Understanding CNNs:
Implement random initialization of weights in a CNN and compare it with a non-
III random initialization; Implement regularization to prevent overfitting in CNNs; 10
Implement different gradient-based approaches for optimization; Implement
efficient gradient computations in CNNs.
CNN Architectures; Applications of CNNs in Computer Vision:
IV Explore the building blocks of state-of-the-art CNN architectures; Explore 10
applications of CNNs to real life problems.
References

1. A Guide to Convolutional Neural Networks for Computer Vision, Salman Khan, Hossein Rahmani, Syed Afaq
Ali Shah, and Mohammed Bennamoun, Morgan & Claypool Publishers, 2018.

2. Lecture slides of Prof. Fei-Fei Li – Stanford University – Available online at http://cs231n.stanford.edu/

3. Neural Networks and Deep Learning, Michael Nielsen, Determination Press – Available online at
http://neuralnetworksanddeeplearning.com/index.html

ME (Artificial Intelligence and Machine Learning) 40

4. Computer Vision: Algorithms and Applications, Richard Szeliski, Springer, 2011 – Online resource from
Springer available at http://szeliski.org/Book/

L T P C Total hours
AML 5283: Natural Language Processing Principles &
Applications Lab
0 0 3 1 36

Course Outcome

1. Develop an in-depth understanding of both algorithms for processing linguistic information and the
underlying computational properties of natural languages.
2. Formulate deep learning approaches for natural language processing tasks.
3. Develop practical experience with state-of-the-art natural language processing tools and libraries.

Unit No. of
Topics
Hours

Introduction to Natural Language Processing (NLP); Regular Expressions; N-gram

Language Models:
I Understand the basics of NLP and role of probability in it; Understand how to use and 14
apply regular expressions; Develop the idea of a probabilistic language model;
Understand how to evaluate and compare language models.

Naive Bayes and Sentiment Classification; Vector Semantics and Embeddings:

II Implement sentiment classification using real-life datasets; Implement building

8
blocks for vector representation of words; Implement and compare vector models for
words; Visualize word embeddings.

NLP with Deep Learning; Applications of Natural Language Processing:

Implement neural models for NLP applications; Gain experience in using PyTorch;
III 14
Implement recurrent neural network models for NLP applications; Explore practical
applications of NLP.

ME (Artificial Intelligence and Machine Learning) 41

References

1. Speech and Language Processing, Dan Jurafsky and James H. Martin, Pearson; 3rd Edition (draft) – Available
online at https://web.stanford.edu/~jurafsky/slp3/

2. Natural Language Processing with Python. – Analysing Text with the Natural Language Toolkit, Steven Bird,
Ewan Klein, and Edward Loper, ISTE Ltd., 1st Edition, 2017 - Available online at
https://www.nltk.org/book/

3. A Primer on Neural Network Models for Natural Language Processing, Yoav Goldberg – Available online at
http://faculty.cse.tamu.edu/huangrh/Spring18/nnlp.pdf

4. Natural Language Processing with PyTorch, Delip Rao & Brian McMahan, O'Reilly, 1st Edition, 2019.

Total
ENP 5280: Entrepreneurship Lab L T P C
hours
0 0 3 1 36

Course Outcome

1. Study of prominence of entrepreneurship.

2. Develop use cases for building a business.
3. Evaluation of factors influencing business venture.

No. of
Unit Topics
Hours
Study of use cases for need and prominence of entrepreneurship, associated decision
I making process. 6

Study of report by the National Knowledge Commission on the importance of

II entrepreneurship in economic development. 6

Develop use cases for identifying and evaluating opportunities, developing business
III plan, assessment of resources, project appraisal and feasibility plan. 9

Creating and starting venture includes legal requirements, marketing strategies,

IV financial plans and human resources management. 9

V Design a Case studies of Indian and International Entrepreneurship. 6

ME (Artificial Intelligence and Machine Learning) 42

References
1. Management and Entrepreneurship, NVR Naidu and T. Krishna Rao, IK International Publishing House
Pvt.Ltd, 2008.
2. Fundamentals of Entrepreneurship, Mohanthy Sangram Keshari, PHI Learning Pvt. Ltd., 2005.

L T P C Total hours
MPT 5200: Mini Project - II
0 0 0 4 48

Course Outcome

1. Identify the real-world and social relevant problems and perform feasibility analysis for finding
solutions.
2. Develop solutions to the identified problems by applying research methodology and development life
cycle with appropriate documentation by incorporating ethical standards.
3. Work effectively as a member in a team and communicate technical information effectively.

No. of
Unit Topics
Hours

I Problem identification, literature survey, formation of detailed specifications.

II Design and implementation of the proposed system architecture. 48

Demonstrate an ability to present and defend project work carried out to a panel
III
of experts.

References

1. Research articles and Online Resources.

ME (Artificial Intelligence and Machine Learning) 43

 L T P C Total hours
PSD 5200: Professional Skill Development - II
0 0 0 1 12

Course Outcome

1. Develop the skills needed for approaching technical and HR interviews.

2. Use mathematical, reasoning, and domain specific skills to solve objective questionnaires in time.
3. Demonstrate depth of knowledge in the chosen field of study.

No. of
Unit Topics
Hours

I Peer interviews, mock interviews.

II Logical reasoning, mathematical aptitude, domain specific problem-solving skills. 12

III Conduction of domain specific knowledge test.

References

1. R S Aggarwal. Quantitative Aptitude for Competitive Examinations. S Chand, 2017.

2. McDowell, Gayle Laakmann. Cracking the coding interview: 189 programming questions and solutions.
CareerCup, LLC, 2015.
3. Domain specific tools and online resources.

L T P C Total hours
AML 6098: Project Work
0 0 0 25 300

Course Outcome

1. Undertake innovative industry/research-oriented projects and perform feasibility analysis for finding
solutions.
2. Implement and test the proposed design using appropriate framework, programming language and tools.
3. Demonstrate an ability to present and defend project work carried out to a panel of experts.

ME (Artificial Intelligence and Machine Learning) 44

 No. of
Unit Topics
Hours

Problem identification, literature survey, formation of detailed requirement

I
specification document.

II Design and implementation of the proposed modules with specific test cases. 300

Detailed report of the work carried out, present, and defend the project work
III
carried out to a panel of experts.

References

1. Research articles, domain specific tools and online resources.

ME (Artificial Intelligence and Machine Learning) 45

 Program Outcome and Course Outcome Mapping

Sl. Course
Course Name Credits PO1 PO2 PO3 PO4 PO5 PO6 PO7 PO8 PO9 PO10 PO11
No. Code

Algorithms and
BDA 5101 Data Structures
1 3 * * * * * * *
for Big Data

Applied Linear
AML 5101
2 3 * * * * * * *
Algebra

Applied Machine
AML 5102 3
3 * * * * * * *
Learning

Applied
AML 5103 Probability &
4 3 * * * * * *
Statistics

Applications of
AML 5131 3
* * * * * *
Graph Theory

Architecture of
BDA 5102
5 3
Big Data Systems

Principles of Data
BDA 5132
3 * * * * *
Visualization

Algorithms and
BDA 5151 Data Structures
6 1 * * * * * * *
for Big Data Lab

AML 5151 Applied Linear

7 Algebra Lab 1 * * * * * *

Applied Machine
AML 5152
8 1 * * * * * * * *
Learning Lab

Applied
AML 5153 Probability &
9 1 * * * * * * *
Statistics Lab

ME (Artificial Intelligence and Machine Learning) 46

 AML 5181
Applications of
1 * * * * * * * * *
Graph Theory Lab

BDA 5152
Architecture of

10 Big Data Systems 1

Lab

BDA 5182
Principles of Data
1 * * * * *
Visualization Lab

MPT 5100 Mini Project – I 4 * * * * * * *

11

Professional Skill * * *
12 PSD 5100 1
Development – I

Advanced
Applications of 3
13 AML 5201 * * * * * * *
Probability &
Statistics

3
14 AML 5202 Deep Learning * * * * * * * *

Machine Learning
3
15 AML 5203 Principles & * * * * * * *
Applications

Reinforcement 3
16 AML 5204 Learning * * * * * *

AML 5231 Applied

Mathematics for * * * * * * *
3
Machine Learning

AML 5232 Convolutional

17
Neural Networks
* * * * * * *
3
for Computer
Vision

AML 5233 Natural Language * * * * * * *

3
Processing

ME (Artificial Intelligence and Machine Learning) 47

 Principles &
Applications

ENP 5230 Entrepreneurship * * * * * *

3

Advanced
18 AML 5251 Applications of * * * * * * *
1
Probability &
Statistics Lab
Deep Learning
19 AML 5252 * * * * * * *
Lab 1

Machine Learning
20 AML 5253 Principles & 1
* * * * * * *
Applications Lab

Reinforcement
21 AML 5254 * * * * *
Learning Lab 1

Applied
Mathematics for
AML 5281 * * * * * *
1
Machine Learning
Lab

AML 5282 Convolutional

Neural Networks
* * * * * *
1
for Computer
22
Vision Lab

AML 5283 Natural Language

Processing
* * * * * * *
1
Principles &
Applications Lab

ENP 5280 Entrepreneurship * * * * * *

1
Lab

23 MPT 5200 Mini Project – II * * * * * * *

4

Professional Skill
24 PSD 5200 1 * * * *
Development – II

25 AML 6098 Project Work 25 *. * * * * * * * * * *

ME (Artificial Intelligence and Machine Learning) 48