Syllabus for
Master of Science (Computational Statistics and Applied AI)
Academic Year  (2024)

MSc (Computational Statistics and Applied AI) is a two year postgraduate programme offered in partnership with CleverInsight, Bangalore. The programme consists of 4 semesters where the classes are conducted on weekends. The industry-integrated MSc programme in Computational Statistics and Applied AI is an excellent opportunity for professionals who want to transition into the data science and AI fields. The programme structure is designed to equip students with industry-focused data science, data engineering and applied AI modules. The programme's practical training and hands-on experience will provide students with the skills and knowledge needed to become data-driven decision-makers. The collaboration with industry experts and live projects provides students with real-world experience, making them industry-ready.

CleverInsight: CleverInsight is a data science company focused on creating meaningful insights from customers' data. With an array of expertise across diverse fields, including banking and finance, healthcare, retail, social media, telecom, IOT, and government, CleverInsight offers an unmatched level of service and support that sets them apart from the competition. Their team of experienced bankers, data scientists, engineers, programmers, and marketers work collaboratively to deliver transformative solutions that help clients optimize their business outcomes, enhance their customer experiences, and promote innovation across their organizations.

CleverInsight takes pride in its ability to stay on the forefront of technological advancements, constantly seeking out cutting-edge tools and strategies to ensure that their clients remain competitive in an ever-changing market. By leveraging the latest technologies and methodologies, CleverInsight empowers clients to maximize their resources, improve operational efficiency, and achieve sustained growth. The company's commitment to excellence is evident in everything they do. From their customer-centric approach to their rigorous attention to detail, CleverInsight delivers measurable results that drive tangible value for their clients. Whether clients are seeking to optimize their business processes, drive innovation, or transform their operations, CleverInsight has the expertise and experience to help them achieve their goals. Website: https://cleverinsight.co

Programme Outcome/Programme Learning Goals/Programme Learning Outcome:

PO2: To identify opportunities in organizations for deploying statistical tools and methods.

PO3: To exercise data driven methodologies using AI and Statistical techniques to enhance organizational performance.

PO4: To exercise data driven methodologies using AI and Statistical techniques to enhance organizational performance.

ESE and CIA - Centeralized for two courses per semester and remaining courses are of type full CIA consisting of theroy and lab assessments (CAC & CAT).

To enable the students to understand the fundamentals of statistics to apply and infer descriptive measures and probability for data analysis. This course will enable students to understand basic probability and distribution, the concept of inference, analysis of variance, and regression and to apply these in real-life scenarios.

Data Representation, Average, Spread, Experiments, Outcomes, Events, Probability, Permutations and Combinations, Random Variables, Probability Distributions, Mean and Variance of a Distribution, Binomial, Poisson, and Hypergeometric Distributions, Normal Distribution, Distributions of Several Random Variables.

Types of Statistical Inference, Descriptive Statistics, Inferential Statistics, Descriptive Statistics, Measures of Central Tendency: Mean, Median, Mode, Midrange, Measures of Dispersion: Range, Variance, Mean Deviation, Standard Deviation. Coefficient of variation: Moments, Skewness, Kurtosis,

Estimation-properties, Methods of estimation-Point estimation, Interval estimation, One sample hypothesis testing: hypothesis, Testing of Hypothesis, Binomial distribution and normal distribution, Chi-Square Tests, t-test, Z test, F- test

Analysis of Variance: Fixed Effects, Random Effects, Mixed Models, 12 Introducing the Analysis of Variance (ANOVA), Performing the ANOVA, Random Effects ANOVA and Mixed Models, One-Way Random Effects ANOVA,

Correlation: Scatter plot, Karl Pearson coefficient of correlation, Spearman's rank correlation coefficient, multiple and partial correlations (for 3 variates only). Simple Linear Regression-OLS estimates- Properties.

[3]. Montgomery D.C and Runger G.C, Applied Statistics and Probability for Engineers, Wiley India, New Delhi, 2013.

[4]. Mood A.M, Graybill F.A and Boes D.C, Introduction to the Theory of Statistics, McGraw Hill, New Delhi, 2008.

ESE - 40%

The objective of this course is to provide comprehensive knowledge of scripting languages required for Data Science. Introduces computational thinking through applications of data science using the Julia programming language.

Structure of Python Program-Underlying mechanism of Module Execution-Branching and Looping-Problem Solving Using Branches and Loops-Functions - Lists and Mutability- Problem Solving Using Lists and Functions

Sequences, Mapping and Sets- Dictionaries- -Classes: Classes and Instances-Inheritance- Exceptional Handling- Introduction to Regular Expressions using “re” module Sequences, Mapping and Sets- Dictionaries- -Classes: Classes and Instances-Inheritance- Exceptional Handling-Introduction to Regular Expressions using “re”module.

Basics of NumPy-Computation on NumPy-Aggregations-Computation on Arrays- Introduction to Pandas Objects-Data indexing and Selection-Operating on Data in Pandas- Handling Missing Data-Hierarchical Indexing - Combining Data Sets- Aggregation and Grouping-Pivot Tables-Vectorized String Operations -Working with Time Series

Introduction, Basic Features, HTML and CSS, Visualizing Data, Managing Data, Creating an interactive dashboard with dc.js-summary

Introduction, Key features - Getting started with Julia- Working with packages and modules- Working with collections in Julia- arrays, Getting the data into a matrix, Computing basic statistics of the data stored in a matrix- Fitting a linear regression- Plotting the Anscombe's quartet data- Handling time series data and missing values- Toolbox for Data Analysis

[2] Jake VanderPlas ,Python Data Science Handbook - Essential Tools for Working with Data, O’Reily Media,Inc

ESE - 40%

Visual Analytics enables effective analytical reasoning based on data and interactive visual interfaces. Exploratory data visualization and analysis, in combination, are used for solving complex problems in data science.

Introduction of visual perception - visual representation of data - Gestalt principles -information overloads - Basic Chart Types- Effectiveness of Data Visualization - Visual Encoding: Marks and Channels - Use of Color in Visualization

Visualization reference model - visual mapping - visual analytics - Design of visualization Applications -Classification of visualization systems - Interaction and visualization techniques misleading. Visualization of one, two and multi-dimensional data, text and text documents Visualization elements -Visualization of groups, trees, graphs, clusters, networks, software - Metaphorical visualization

Design Principles- Choosing the Right Chart Type- Creation-formatting-create a device preview - filters - objects - story creation

Visual Analytics for Data Exploration- - Visual Analytics for Model Building and Understanding - Visual Analytics Techniques

Visualization of volumetric data, - vector fields, processes and simulations - Visualization of maps, geograp- Online Visualization Tools- Plotly, Sisense, PowerBI, IBM Watson analytics, Kibana, Grafana, D3.js, Fusion charts, Tableau public, charted, Google charts, Flot, chartist.js, High charts, Datawrapper, Dygraphs, Raw, Timeline JS, Polymaps

[2] Visualization Analysis & Design ,Tamara Munzner and Eamonn Maguire, A K Peters/CRC Press,2014 ISBN: 978-1-4665-0891-0

ESE - 40%

This course is designed to provide a comprehensive understanding of Prescriptive and Operational Models to individuals with little to no experience in data science, as well as to working professionals who want to expand their knowledge and skills in these areas. The course covers the basics of prescriptive modeling, such as optimization and decision analysis, and operational models, such as forecasting, simulation, and risk analysis.

Overview of Prescriptive and Operational Models -Linear Programming-Sensitivity Analysis and Interpretation

Non-linear Programming-Decision Analysis -Decision Tree Modeling

Forecasting-Time Series Analysis-ARIMA and Seasonal ARIMA models

Simulation-Monte Carlo Simulation-Risk Analysis

Optimization Techniques in Operational Models-Transportation and Assignment Problems-Case studies and real-world applications.

[2] Dursun Delen, “Prescriptive Analytics: The Final Frontier for Evidence-Based Management and Optimal Decision Making”, Pearson FT Press; 1st edition (24 October 2019)

[2] Dursun Delen, “Prescriptive Analytics: The Final Frontier for Evidence-Based Management and Optimal Decision Making”, Pearson FT Press; 1st edition (24 October 2019)

ESE - 40%

The course covers the topics of Data Engineering, Data pipelines, and big data technologies. It focuses on the implementation of Big Data Technologies such as Hadoop, Waddle, Spark, and Kafka and the application of Data pipeline tools.

Overview of Data Engineering and its importance -Data Modeling and Data Architecture -Understanding of Data Warehousing and ETL Processes.

Data Storage and Retrieval-SQL and NoSQL databases-Data warehousing, data marts, and data lakes.

Data Pipeline Architecture and Design-Batch Processing and Real-time processing-ETL (Extract, Transform, Load) and ELT (Extract, Load, Transform) processes.

Introduction to Big Data and its challenges-Distributed Computing and Parallel Processing-Big Data Technologies such as Hadoop, Waddle, Spark, and Kafka.

Data Pipeline Tools such as Apache Airflow, Luigi, or AWS Glue-Deployment and Maintenance of Data Pipelines-Best Practices and Case Studies.

ESE - 40%

This course enables the students to apply the concepts of neural networks and understand the concepts and models of the neural networks and deep learning and its applications.

Neural Networks-Application Scope of Neural Network Fundamental Concept of ANN: The Artificial Neura Network Biological Neural Network-Comparison between Biological Neuron and Artificial Neuron-Evolution of Neural Network

Basic models of ANN-Learning Methods- Activatio Functions Importance Terminologies of ANN

Shallow neural networks- Perceptron Networks-Theory-Perceptron Learning RuleArchitecture-Back Propagation Network- Theory - Architecture - Learning Factors for Back-Propagation Network.

Introduction - Components of CNN Architecture - Rectified Linear Unit (ReLU) Layer - -Architectures of CNN - Applications of CNN 

Introduction- The Architecture of Recurrent Neural Network- Echo-State Networks- Long Short-Term Memory (LSTM) - Applications of RNN

Introduction - Features of Auto encoder Types of Autoencoder Restricted Boltzmann Machine- Boltzmann Machine

[2] Dr. S Lovelyn Rose, Dr. L Ashok Kumar, Dr. D Karthika Renuka, Deep Learning Using Python, Wiley-India, 1st Edition, 2019.

[2] Francois Chollet, Deep Learning with Python, Manning Publications; 1st edition, 2017

[3] John D. Kelleher, Deep Learning (MIT Press Essential Knowledge series), The MIT Press, 2019

The goal is to make familiar with the concepts of the study of human language from a computational perspective. It  covers syntactic, semantic and discourse processing models, emphasizing machine learning concepts.

Introduction to NLP- Background and overview- NLP Applications -NLP hard AmbiguityAlgorithms and models, Knowledge Bottlenecks in NLPIntroduction to NLTK, Case study.

Word Level Analysis: Regular Expressions, Text Normalization, Edit Distance, Parsing and Syntax- Spelling, Error Detection and Correction-Words and Word classesPart-of Speech Tagging, Naive Bayes and Sentiment Classification: Case study.

Semantic Analysis: Meaning Representation-Lexical Semantics- Ambiguity-Word Sense Disambiguation. Discourse Processing: cohesion-Reference Resolution- Discourse Coherence and Structure.

N-gram Language Models: N-Grams, Evaluating Language Models -The language modeling problem

Natural Language Generation: Architecture of NLG Systems, Applications.

MachineTranslation: Problems in Machine Translation-Machine Translation Approaches- Evaluation of Machine  Translation systems.

Case study: Characteristics of Indian Languages.

Information Retrieval: Design features of Information Retrieval Systems-Classical, Non-classical, Alternative Models of Information Retrieval – valuation Lexical 

Resources: Word Embeddings - Word2vec-Glove.

Unsupervised Methods In NLP Graphical Models for Sequence Labelling in NL

[2] Daniel Jurafsky and James H., Speech and Language Processing, 2nd Edition, Martin Prentice Hall, 2013.

[2] Steven Bird, Ewan Klein and Edward Loper, Natural Language Processing with Python, O’Reilly Media; First edition, 2009

The Effective Data Driven Decision Science program is designed to equip students with the skills necessary to leverage data and analytics in decision making. Through this program, students will develop a deep understanding of data collection, analysis, and visualization techniques, as well as an understanding of how to develop compelling data narratives that drive effective decision-making.

Introduction to data-driven decision making and the role of  data in decision making - Key concepts and terminology in data science, including.

Descriptive, predictive, and prescriptive analytics - Overview of the data science process and how it can be applied to decision making - BUS framework for analytics story framing.

Techniques for collecting and cleaning data to ensure accuracy and completeness - Data preprocessing techniques such as data normalization, feature scaling, and dimensionality reduction - Exploratory data analysis techniques to identify patterns and trends in data. 

Techniques for developing predictive models using machine learning algorithms - Supervised learning techniques such as regression and classification - Unsupervised learning techniques such as clustering and anomaly detection.

Overview of prescriptive analytics and its role in decision making - Techniques for developing prescriptive models using optimization and simulation techniques - Overview of decision support systems and how they can be used to facilitate decision making. 

Techniques for implementing data-driven decisions in organizations - Effective communication strategies for presenting data and insights to decision makers - Ethical  considerations in data-driven decision making and the importance of data privacy and security 

[2] Cole Nussbaumer Knaflic, “Storytelling with Data: A Data Visualization Guide for Business Professionals", Wiley; 1st edition (20 November 2015) 

The course covers the topics of Applied AI in Business, BFSI, Healthcare, Bioinformatics and Manufacturing

Overview of AI in Business - Importance of AI in BFSI, Healthcare, Bioinformatics, and Manufacturing - Impact of AI on Business Processes and Decision Making - Case Studies of Successful Implementation of AI in Business

Introduction to BFSI Industry - Applications of AI in BFSI including Fraud Detection, Risk Management, Customer Service, and Investment Management - Regulatory and Ethical Considerations in AI Applications in BFSI - Realworld Case Studies of AI Implementation in BFSI.

Introduction to Healthcare Industry - Applications of AI in Healthcare including Diagnostics, Drug Discovery, Personalized Medicine, and Medical Imaging - Ethical and Regulatory Considerations in AI Applications in Healthcare - Real-world Case Studies of AI Implementation in Healthcare

Introduction to Bioinformatics - Applications of AI in Bioinformatics including DNA Sequencing, Protein Folding, and Drug Design - Challenges in AI Applications in Bioinformatics - Real-world Case Studies of AI Implementation in Bioinformatics 

Introduction to Manufacturing Industry - Applications of AI in Manufacturing including Predictive Maintenance, Quality Control, and Supply Chain Management - Ethical and Regulatory Considerations in AI Applications in Manufacturing - Real-world Case Studies of AI Implementation in Manufacturing.

[2] Alex Castrounis, “AI for People and Business: A Framework for Better Human Experiences and Business Success”, O′Reilly (23 July 2019)

[3] Raj Singh, “Artificial Intelligence In Banking & Finance: How AI is Impacting the Dynamics of Financial Services”, Adhyyan Books (14 January 2019)

The explosive growth of the “Internet of Things” is changing our world and the rapid growth of IoT components is allowing people to innovate new designs and products at home. Explore the basics of cloud analytics and the major cloud solutions and learn how to analyze extremely large data sets, and to create visual representations of that data

Introduction to IoT - Definition and Characteristics, Physical Design Things- Protocols, Logical Design- Functional Blocks, Communication Models- Communication APIsIntroductiontomeasurethephysicalquantities, IoT Enabling Technologies - WirelessSensor Networks, Cloud Computing Big Data Analytics, Communication Protocols- Embedded System- IoT Levels and Deployment Templates. 

Introduction to Smart Systems using IoT - IoT Design Methodology- IoT Boards (Raspberry Pi, Arduino) and IDE.

CaseStudy: Weather Monitoring - Logical Design using Python, Data types & Data Structures- Control Flow, Functions Modules- Packages, File Handling - Date/Time Operations,

Sensing and Sensors-Wireless Sensor Networks, Challenges and Constraints-Applications: Structural Health Monitoring,

Traffic Control, Health Care - Node Architecture - Operating system. Introduction – Fundamentals of MAC Protocols – MAC protocols for WSN – Sensor MAC Case Study –Routing Challenges and Design Issues – Routing Strategies– Transport Control Protocols–Transport Protocol Design Issues– Performance of Transport Protocols

Major benefits of cloud computing - Cloud computing deployment models - Types of cloud computing services - Emerging cloud technologies and services - Risks and challenges with the cloud - Parameters before adopting cloud strategy - Technologies utilized by cloud computing. Amazon Relational Data Store - Amazon DynamoDB - Google Cloud SQL - Google Cloud Datastore - Windows Azure SQL Database - Windows Azure Table Service

Cloud Dataflow - The Dataflow programming model - Cloud Pub/Sub - Cloud storage - Cloud SQL - Cloud BigTable - Cloud Spanner - Cloud Datastore - Persistent disks . Cloud Natural Language API – TensorFlow - Cloud Speech API - Cloud Translation API - Cloud Vision API - Cloud Video Intelligence – Dialogflow – AutoML

[2] Kazem Sohraby, Daniel Minoli and TaiebZnati, Wireless Sensor Networks: Technology. Protocols and Application, Wiley Publications, 2010.

[3] Waltenegus Dargie and Christian Poellabauer, Fundamentals of Wireless Sensor Networks: Theory and Practice, A John Wiley and Sons Ltd., 2010. 

[4] Snket Thodge, Cloud Analytics with Google Cloud Platform, Packt Publishing, 2018. 

[5] Arshdeep Bahga and Vijay Madisetti, Cloud computing - A Hands-On Approach, Create Space Independent Publishing Platform, 2014.

[2] Michael Miller, The Internet of Things, Pearson Education, 2015. 

[3] Holger Karl and Andreas Willig, Protocols and Architectures for Wireless Sensor Networks, John Wiley & Sons Inc., 2005.

[4] Erdal Çayırcı and Chunming Rong, SecurityinWirelessAdHocandSensorNetworks, John Wiley and Sons, 2009.

[5] Carlos De Morais Cordeiro and Dharma Prakash Agrawal, Ad Hoc and Sensor Networks: Theory and Applications, World Scientific Publishing, 2011.

[6] Adrian Perrig and J.D.Tygar, Secure Broadcast Communication: In Wired and Wireless Networks, Springer, 2006.

[7] Deven Shah, Kailash Jayaswal, Donald J. Houde, Jagannath Kallakurchi, Cloud Computing - Black Book, Wiley, 2014.

[8] Thomas Erl, Ricardo Puttini, Zaigham Mahmood, Cloud Computing: Concepts, Technology & Architecture, Prentice Hall, 2014

This course aims to equip students with the fundamental knowledge and skills required for real-time data analysis and time series modeling. It covers the basics of time series data, stochastic processes, and various modeling techniques, enabling students to analyze data, make forecasts, and understand the dynamics of time-dependent phenomena.

Overview of real-time data sources and applications - Understanding the challenges and opportunities in real-time data analysis – Basics of time series data and its characteristics - Stochastic Process – The Autocovariance and autocorrelation function – The partial autocorrelation function – White noise process – Estimation of Sample Mean, Autocovariance and Autocorrelation.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Autoregressive Process – First order AR process – Second order AR process – General pth order AR process – Moving Average Process – First order MA process – Second order MA process – General qth order MA process – Dual relationship between AR(p) and MA(q) processes – Autoregressive Moving Average ARMA (p, q) processes.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class ActivityA 

NON-STATIONARY TIME SERIES

Non-stationarity in Mean – Deterministic Trend Models – Stochastic Trend Models and Differencing – Autoregressive Integrated Moving Average (ARIMA) Models – General ARIMA Model – Random Walk model – Non-stationarity in the variance and autocovariance – Variance Stabilizing transformation.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Analysis of seasonal models - parsimonious models for  seasonal time series - Seasonal unit root test (HEGY test) - General multiplicative seasonal models - Seasonal ARIMA  models - estimation - Residual analysis for seasonal time  series. 

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

In sample and out of sample forecast - Simple exponential and  moving average smoothing - Holt Exponential Smoothing - Winter exponential smoothing - Forecasting trend and seasonality in Box Jenkins model:- Method of minimum mean squared error (MMSE) forecast - their properties – forecast error

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] William W. S Wei., Time Series Analysis Univariate and Multivariate Methods. Pearson Education, 2006. 

[2] Brockwell, P. J., & Davis, R. A., Introduction to time series and forecasting. Springer, 2016. 

ESE - 40%

This course will help the students to understand the fundamentals of deep learning, including artificial neural networks (ANNs), perceptrons, activation functions, and training techniques such as gradient descent and backpropagation. It helps to explore Convolutional Neural Networks (CNNs), including their architecture, components, advanced techniques, and applications in image processing tasks. This course will give the knowledge of Generative Adversarial Networks (GANs), their architecture, training methods, variants, and applications in image generation and translation.

FUNDAMENTALS OF DEEP LEARNING

Introduction to Deep Learning

Introduction to artificial neural networks (ANNs) - Perceptrons and activation functions - Training neural networks: gradient descent and backpropagation

Deep Learning Basics

Introduction to deep learning concepts - Basics of optimization algorithms (SGD, Adam, etc.) - Regularization techniques (dropout, weight decay)

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

CNN Architecture and Components

CNN architecture overview - Convolutional layers, pooling layers, activation functions - Common CNN architectures: LeNet, AlexNet, VGG, etc.

Advanced CNN Techniques

Transfer learning and fine-tuning - Object detection with CNNs: R-CNN, Fast R-CNN, Faster R-CNN - Semantic segmentation with CNNs: U-Net, FCN

CNN Applications

Image classification - Object detection and localization - Image segmentation

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity 

RNN Basics

Introduction to recurrent neural networks - Vanishing and exploding gradients problem - Long Short-Term Memory (LSTM) networks - Gated Recurrent Units (GRUs)

Advanced RNN Techniques

Bidirectional RNNs - Attention mechanisms in RNNs - Sequence-to-sequence models - Topic 8: RNN Applications - Language modeling and text generation - Machine translation - Time series prediction

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Introduction to GANs

Basics of Generative Adversarial Networks - GAN architecture: generator and discriminator - Training GANs: adversarial training

Advanced GAN Techniques and Applications

Variants of GANs: DCGAN, WGAN, CycleGAN - GAN applications: image generation, style transfer, image-to-image translation

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Introduction to Reinforcement Learning

Basics of reinforcement learning - Markov Decision Processes (MDPs) - Q-learning and policy gradient methods

Deep Reinforcement Learning

Deep Q-Networks (DQN) - Policy gradient methods with neural networks - Actor-Critic methods

Reinforcement Learning Applications

Game playing: AlphaGo - Robotics and control - Autonomous agents

 Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Charu C. Aggarwal, "Neural Networks and Deep Learning: A Textbook", Springer International Publishing AG; 1st ed., 2018.

[3] François Chollet, "Deep Learning with Python", Manning, First Edition, 2017.

[4] Delip Rao and Brian McMahan, "Natural Language Processing with PyTorch", O′Reilly, 2019.

[5] Ian Goodfellow et al, "Generative Adversarial Networks"

[6] Jakub Langr and Vladimir Bok, "GANs in Action: Deep Learning with Generative Adversarial Networks",  Manning; First Edition, 2019.

[7] Richard S. Sutton and Andrew G. Barto, "Reinforcement Learning: An Introduction", MIT Press; Second edition, 2018.  

[8] Maxim Lapan, "Deep Reinforcement Learning Hands-On", Packt Publishing, 2018.

[2] Ivan Vasilev and Daniel Slater, "Python Deep Learning", Packt Publishing; 2nd edition, 2019.

[3] Aurélien Géron, "Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow", Shroff/O'Reilly, Third edition, 2022.

[4] Mohamed Elgendy, "Deep Learning for Vision Systems", Manning; 1st edition, 2020.

[5] Palash Goyal et al., "Deep Learning for Natural Language Processing", Apress; 1st edition, 2018.

[6] Sean Saito, "Python Reinforcement Learning Projects", Packt Publishing, 2018.

[7] Andrea Lonza, "Reinforcement Learning Algorithms with Python", Packt Publishing, 2019.

ESE - 40%

Multivariate data analysis is introduced in this course. The students' exposure to multivariate data structures, multinomial and multivariate normal distributions, parameter estimation and testing, and a variety of data reduction techniques will aid in their comprehension of research data, presentation, and analysis.

Categorical response data - Probability distributions for categorical data - Statistical inference for discrete data

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Probability    structure    for    contingency    tables    - Comparing proportions with 2x2 tables - The odds ratio -  Tests for independence - Exact inference

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Components of a generalized linear model - GLM for binary and count data - Statistical inference and model checking - Fitting GLMs

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Interpreting the logistic regression model - Inference for logistic regression - Logistic regression with categorical predictors - Multiple logistic regression - Summarizing effects - Building and applying logistic regression models

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Loglinear models for two-way and three-way tables - Inference for Loglinear models - the log-linear-logistic connection - Independence graphs and collapsibility – Models for matched pairs: Comparing dependent proportions.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Agresti, A. Analysis of ordinal categorical. John Wiley & Sons, 2010.

[3] Stokes, M. E., Davis, C. S., & Koch, G. G., Categorical data analysis using SAS. SAS Institute., 2012.

[4] Agresti, A., An introduction to categorical data analysis. John Wiley & Sons, 2018.

[5] Bilder, C. R., & Loughin, T. M., Analysis of categorical data with R., Chapman and Hall/CRC, 2014.

ESE - 40%

Multivariate data analysis is introduced in this course. The students' exposure to multivariate data structures, multinomial and multivariate normal distributions, parameter estimation and testing, and a variety of data reduction techniques will aid in their comprehension of research data, presentation, and analysis.

Basic concepts on the multivariate variable and representations of random vectors, mean vectors, covariance matrices - Bivariate normal distribution; an overview. Multivariate normal distribution and its properties, Its expectation, and Variance-Covariance matrix. Conditional distributions and Independence of random vectors. Multinomial Distribution

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Maximum likelihood estimation of mean vector and var-covariance matrix-Sampling distribution of mean vector and var-covariance matrix -  Wishart Distribution - Tests of hypotheses about the mean vectors and covariance matrices for multivariate normal populations - Hotellings T2 - Comparing Mean Vectors from Two Populations - Mahalanobis Distance-Box-M Test.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Multivariate analysis of variance (MANOVA) of one and two-way classified data with their interactions - Univariate and Multivariate Two-Way Fixed-effects Model with Interaction. Concept of MANCOVA.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Concepts of discriminant analysis - Computation of linear discriminant function (LDF) - Classification between k multivariate normal populations based on LDF - Fisher’s Method for discriminating two or several populations - Evaluating Classification Functions - Probabilities of misclassification and their estimation

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Principal components, sample principal components asymptotic properties. Canonical variables and canonical correlations: definition, estimation, computations. Factor analysis: Orthogonal factor model, factor loadings, estimation of factor loadings.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Srivastava, M.S. and Khatri, C.G., An Introduction to Multivariate Statistics. North Holland., 1979.

[3] Muirhead, R.J. Aspects of Multivariate Statistical Theory. John Wiley. New York., 1982.

[4] Rohatgi, V.K. and Saleh, A.K.M.E., An Introduction to Probability Theory and Mathematical Statistics. 2nd edn. John Wiley & Sons. New York, 2015.

ESE - 40%

This course offers an in-depth exploration of causal inference, a critical aspect of scientific inquiry across various disciplines. Through lectures, readings, and practical exercises, students delve into fundamental concepts, methodologies, and advanced topics in causal inference. The course covers the theoretical underpinnings of causality, methods for estimating causal effects from observational and experimental data, as well as contemporary issues and applications in public health, economics, and social sciences. Students engage in project work where they apply causal inference techniques to real-world datasets, fostering critical thinking and analytical skills.

Fundamentals of Causal Inference

Definition of causality and causal inference -Types of causal relationships: deterministic vs. probabilistic, direct vs. indirect - Counterfactual framework and potential outcomes

Causal Diagrams and Graphical Models

Directed Acyclic Graphs (DAGs) as a tool for causal inference - Understanding causal relationships using causal diagrams - Causal identification and estimation using graphical models

Assumptions and Identification of Causal Effects

Assumptions of causal inference: consistency, exchangeability, positivity, no hidden confounding - Identifiability conditions for causal effects

Methods for assessing and Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Observational Studies and Causal Inference

Challenges of causal inference in observational studies - Methods for estimating causal effects from observational data: propensity score matching, inverse probability weighting, instrumental variables

Experimental Design and Randomized Controlled Trials (RCTs)

Introduction to experimental design and RCTs - Randomization and its role in causal inference - Estimating causal effects in RCTs: intention-to-treat analysis, per-protocol analysis

Quasi-Experimental Designs and Natural Experiments

Quasi-experimental designs: difference-in-differences, regression discontinuity - Identifying causal effects using natural experiments and instrumental variables - Limitations and considerations in quasi-experimental designs

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity 

Mediation and Moderation Analysis

Understanding mediation and moderation in causal inference - Methods for assessing and estimating mediation effects: causal mediation analysis - Interpreting moderation effects and interactions

Time-Series Analysis and Causal Inference

Time-series data and its relevance to causal inference - Granger causality and its limitations - Causal inference in time-series analysis: structural equation modeling, dynamic causal modeling

Machine Learning and Causal Inference

Introduction to causal machine learning - Propensity score-based methods for machine learning - Challenges and considerations in using machine learning for causal inference

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Causal Inference in Public Health

Applications of causal inference in public health research - Case studies in causal inference applied to epidemiology and health policy - Ethical considerations in causal inference research

Causal Inference in Economics

Economic applications of causal inference: labor economics, development economics, policy evaluation - Case studies and empirical methods in economic causal inference - Policy implications and decision-making based on causal analysis

Causal Inference in Social Sciences

Causal inference methods in social science research - Case studies and examples from sociology, political science, and psychology - Challenges and future directions in causal inference in social sciences

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Project Development and Presentations

●      Students work on a project applying causal inference methods to real-world data

●      Designing and executing a research question

●      Project presentations and feedback sessions

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Stephen L. Morgan and Christopher Winship, "Counterfactuals and Causal Inference: Methods and Principles for Social Research", Cambridge University Press; 1st edition 2007.

[3] Paul R. Rosenbaum, "Observational Studies", Springer; 2nd edition, 2002.

[4] Paul R. Rosenbaum, "Design of Observational Studies", Springer-Verlag New York Inc., 2012.

[5] Miguel A. Hernán and James M. Robins, "Causal Inference: What If", CRC Press Inc; 1st edition, 2023.

[6] Judea Pearl, "Causal Inference in Statistics: A Primer", Wiley, 1st edition, 2016.

[7] Guido W. Imbens and Donald B. Rubin, "Causal Inference for Statistics, Social, and Biomedical Sciences: An Introduction", Cambridge University Press, 1st edition, 2015.

[2] Joshua D. Angrist and Jörn-Steffen Pischke, "Mostly Harmless Econometrics: An Empiricist's Companion", Princeton University Press, 1st edition, 2009.  

[3] Judith D. Singer and John B. Willett, "Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence", OUP USA; 1st edition, 2003.

[4] Jasjeet S. Sekhon, "Applied Causal Analysis"

[6] Guido W. Imbens and Donald B. Rubin, "Regression Analysis and Causal Inference: A Second Course in Statistics"

ESE - 40%

This course provides an in-depth exploration of No Code Analytics Platforms, focusing on empowering individuals with the ability to analyze data, build dashboards, deploy machine learning models, automate workflows, and collaborate effectively without the need for extensive programming knowledge. Students will delve into the landscape of No Code development, examining its advantages, challenges, and applications in analytics and data science. Through hands-on exercises and project-based learning, students will gain practical experience in utilizing leading No Code tools to solve real-world analytics problems.

Overview of No Code Development

Introduction to No Code and Low Code development - Advantages and challenges of No Code platforms - Use cases and applications in analytics and data science

No Code Analytics Platform Landscape

Review of leading No Code analytics platforms - Comparison of features and capabilities - Understanding data connectors and integrations

Data Preparation and Integration in No Code Platforms

Importing and connecting data sources - Data cleaning and transformation workflows - Introduction to data modeling and schema design

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Introduction to Dashboard Design

Principles of effective dashboard design - Visualization best practices - Selecting appropriate chart types and widgets

No Code Dashboard Builders

Exploring popular dashboard building tools - Creating interactive dashboards without code - Customizing layouts, themes, and styles

Advanced Dashboard Features

Implementing filters and drill-downs - Adding interactivity with actions and events - Real-time data updates and streaming analytics

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity 

Introduction to Automated Analytics

Overview of automated insights and analysis - Leveraging AI and machine learning in No Code platforms - Understanding automated data exploration techniques

No Code Machine Learning Models

Building predictive models without code - Model training and evaluation workflows - Deploying and integrating machine learning models in applications

Model Interpretability and Explainability

Explaining machine learning predictions - Interpreting model outputs and insights - Ensuring transparency and accountability in AI-driven decisions

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

No Code Workflow Automation

Introduction to workflow automation tools - Creating and managing automated workflows - Integrating with third-party services and APIs

Collaboration and Teamwork

Collaborative features in No Code platforms - Version control and sharing functionalities - Best practices for team collaboration in analytics projects

Governance and Compliance

Ensuring data security and privacy - Compliance with regulatory requirements - Establishing governance policies and procedures

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Project Development and Presentations

●      Students work on a project using a No Code analytics platform

●      Designing, implementing, and iterating on analytics solutions

●      Project presentations and feedback sessions

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Bram Kanstein and Lacey Kesler, "The No Code MBA"

[3] Stephen Few, "Information Dashboard Design",  Analytics Press; Second edition, 2013.

[4] Cole Nussbaumer Knaflic, "Storytelling with Data", Wiley; 1st edition, 2015.

[5] Frank Hutter et al., "Automated Machine Learning: Methods, Systems, Challenges", Springer, 1st ed.,  2019.

[6] Christoph Molnar,  "Interpretable Machine Learning", Lulu.com, 2020.

[7] Ken Hess, "Workflow Automation: The Key to Modern Business"

[8] Ralph Kimball and Margy Ross, "The Data Warehouse Toolkit: The Definitive Guide to Dimensional Modeling", John Wiley & Sons, 3rd edition, 2013.

[2] Alex Salazar, "No-Code: The Future of Web Development"

[3] Nathan Yau , "Data Points: Visualization That Means Something", John Wiley & Sons Inc, 1st edition, 2013.

[4] Emmanuel Ameisen, "Building Machine Learning Powered Applications", O′Reilly, 2020.

[5] Tonio Buonassisi and Arun Verma, "Explainable AI: Interpreting, Explaining and Visualizing Deep Learning".

[6] Alistair Croll and Benjamin Yoskovitz, "Lean Analytics: Use Data to Build a Better Startup Faster", O'Reilly Media, 1st edition, 2013.

[7] Bruce Schneier, "Data and Goliath: The Hidden Battles to Collect Your Data and Control Your World", W. W. Norton & Company, 2016.

ESE - 40%

This course is designed to provide students with the opportunity to integrate and apply the knowledge, skills, and competencies acquired throughout their academic journey. This course serves as the culmination of the student's academic experience, enabling them to tackle real-world problems, explore innovative solutions, and demonstrate their readiness for professional endeavors. Through a blend of research, analysis, critical thinking, and practical application, students will engage in a comprehensive project that reflects their chosen field of study or specialization.

This project is to be taken up either in the industry or in an R&D organization. The students will engage in a comprehensive project that reflects their chosen field of study or specialization.

ESE - 40%

This course provides an in-depth exploration of two critical areas in artificial intelligence (AI): explainability and generative modeling. Students will delve into the foundations of AI explainability, understanding the importance and motivations behind transparent AI systems. The course will cover various techniques and models used for interpretability and explanation, including linear models, decision trees, and post-hoc explainability techniques. The course culminates in a project where students apply their knowledge to develop and present a project related to explainable and generative AI, fostering practical skills in implementation, experimentation, and evaluation.

Introduction to Explainable AI

Motivation and importance of AI explainability - Types of explainability: model-level, instance-level, feature-level - Interpretability vs. Explainability

Interpretable Models and Techniques - Linear models and their interpretability - Decision trees and rule-based models - Feature importance techniques: SHAP, LIME, etc.

Post-hoc Explainability Techniques

Surrogate models - Sensitivity analysis - Local explanation methods

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Introduction to Generative Models

Basics of generative modeling - Probabilistic modeling vs. Deep generative models - Types of generative models: autoregressive, variational autoencoders (VAEs), generative adversarial networks (GANs)

Variational Autoencoders (VAEs)

Architecture and working principles - Learning latent space representations - Applications of VAEs

Generative Adversarial Networks (GANs)

GAN architecture and components - Training GANs: adversarial training - Applications of GANs: image generation, style transfer, data augmentation

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity 

Explainability in Generative Models

Challenges of explaining generative models - Interpretability of latent spaces - Interpreting and explaining GANs outputs

Interpretability Techniques for Generative Models

Activation maximization - Feature visualization - Latent space analysis

Case Studies and Applications

Explainable and interpretable generative models in healthcare - Ethics and fairness in generative AI - Regulatory considerations

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Advanced Generative Models

Adversarial examples and robustness in GANs - Self-attention mechanisms in generative models - Energy-based models and their interpretability

Uncertainty Estimation in Generative Models

Bayesian deep learning and uncertainty quantification - Probabilistic generative models - Evaluating uncertainty in generated samples

Novel Applications and Emerging Trends

Explainable AI for reinforcement learning and robotics - Future directions and emerging research trends in explainable and generative AI - Collaborative and interdisciplinary approaches to advancing the field

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Project Development and Presentations

●      Students work on a project related to explainable and generative AI

●      Implementation, experimentation, and evaluation

●      Project presentations and feedback sessions

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2]  Wojciech Samek (Editor), Grégoire Montavon (Editor), Andrea Vedaldi (Editor), Lars Kai Hansen (Editor), " Explainable AI: Interpreting, Explaining and Visualizing Deep Learning”, Springer Nature Switzerland AG; 1st ed. 2019.

[3] Ian Goodfellow, Yoshua Bengio, and Aaron Courville, "Deep Learning" (Chapter 20)

[4] David Foster, "Generative Deep Learning", O′Reilly, 2019.

[5]  Henning Petzka and Christoph Molnar,  "Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models"

[6] Christoph Molnar, "Interpretable Machine Learning: A Guide for Making Black Box Models Explainable".

[7] Ke Li et al, "Generative Models for Realistic Image Synthesis and Domain Adaptation".

[8] Marzyeh Ghassemi and Tristan Naumann, "Interpretable Machine Learning for Healthcare".

[2] Michael Tober, "Explainable AI: A Guide to Understanding, Visualizing and Interpreting Deep Learning Models"

[3]  John Hany, "Hands-On Generative Adversarial Networks with PyTorch", Packt Publishing 2019.

[4] Rajalingappaa Shanmugamani,  "Deep Learning for Computer Vision", Packt Publishing; 1st Edition, 2018. 

[5] Sandra Wachter, Brent Mittelstadt, and Chris Russell, "Explainable AI: A Visual Guide"

[6] Eamonn Keogh et al., "Interpretable AI: Explainable AI and the Interpretability Frontier"

[7] Anthony D. Joseph et al., "Adversarial Machine Learning",  Cambridge University Press, 2019.

[8] Yarin Gal, "Uncertainty in Deep Learning".

The main objective of this course is to teach students how to define reinforcement learning problems and apply algorithms such as dynamic programming, Monte Carlo, and temporal-difference learning to solve them. Students will advance towards more complex state space environments by employing function approximation, deep Q-networks, and cutting-edge policy gradient techniques. We will also discuss current approaches rooted in reinforcement learning, including imitation learning, meta learning, and more intricate environment formulations.

Reinforcement Learning , Elements of Reinforcement Learning ,An Extended Example: Tic-Tac-Toe, The Agent-Environment Interface , Goals and Rewards ,Unified Notation for Episodic and Continuing Tasks ,The Markov Property ,Markov Decision Processes,Value Functions ,Optimal Value Functions ,Optimality and Approximation.

Lab Exercise

1.     Hands-on activity: participants formalize a simple problem as an MDP.

2.     Familiarize students with the OpenAI Gym library for reinforcement learning.

3.     Set up OpenAI Gym and create a simple environment

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Dynamic Programming Approaches

Policy Evaluation ,Policy Improvement ,Asynchronous Dynamic Programming ,Generalized Policy Iteration ,Monte Carlo Methods - Monte Carlo Policy Evaluation- Monte Carlo Estimation of Action Values- Monte Carlo Control- On-Policy Monte Carlo Control- Off-Policy Monte Carlo Control, Temporal-Difference Learning- TD Prediction

-Advantages of TD Prediction Methods,

Lab Exercises 

1.     Design and implement a Monte Carlo algorithm for solving a specific environment, and discuss the key components of the algorithm, such as episode generation, state-value estimation, and policy improvement.

2.     Implement a Q-learning algorithm for a discrete environment.

3.     Implement the Temporal Difference (TD) prediction algorithm (e.g. Q-learning)

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity 

Exploration Strategies - Epsilon-greedy   - Softmax exploration   - Upper Confidence Bound (UCB)

Multi-Armed Bandits - Contextual bandits   - Thompson Sampling

Adversarial Bandits - Adversarial bandit settings -  Regret minimization

Lab Exercises

1.     Implementing an Adversarial Bandit algorithm in python framework.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Q-networks & Policy optimization

Q-Learning with Neural Networks - Deep Q Networks (DQN) -   - Experience replay - Target networks

Policy Gradients - Policy parameterization - REINFORCE algorithm

Actor-Critic Methods - Advantage functions - A3C (Asynchronous Advantage Actor-Critic)

Lab  Exercises

1.     Implementation exercise using a RL library (e.g., TensorFlow or PyTorch)

2.     Building a simple Actor-Critic model for a basic environment (e.g., CartPole)

3.     Implement a state-of-the-art policy optimization algorithm.

4.     Implement an Actor-Critic algorithm for continuous action space

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

TD-Gammon - Samuel's Checkers Player - The Acrobot - Elevator Dispatching - Dynamic Channel Allocation - Job-Shop Scheduling

Lab Exercises 

1.     Implementation of DQN for the chosen game environment

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Dimitri Bertsekas and John G. Tsitsiklis, Neuro Dynamic Programming, Athena Scientific. 1996.

[2] Deep Learning. Ian Goodfellow and Yoshua Bengio and Aaron Courville. MIT Press. 2016.

ESE - 40%

This course is designed to introduce the concepts of Stochastic Processes. In particular, memoryless phenomena are often encountered in practice, where the future state of a system only depends on its present state, with no recollection of the past. In these cases, Markov chain models offer a formidable tool to answer questions related to the probability of events of practical interest in business and industry. This course provides an overview of concept of Markov Chains and Poisson Processes. It also provides knowledge about the applications of stochastic processes in queueing systems and other simulation studies.

Classification of Stochastic Processes, Markov Processes—Markov Chain-Countable State Markov  Chain. Transition Probabilities, Chapman –  Kolmogorov's Equations, Calculation of n-step Transition Probability and its limit. Classification of

States, Recurrent and Transient States-

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Counting Process-Poisson process and its properties – Interarrival and waiting time distributions. Continuous time Markov chains – Birth and Death Processes, Yule Furry Processes- Kolmogorov's Differential Equations, Applications

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Branching Processes —Galton—Watson Branching Process- Properties of Generating Functions— Extinction Probabilities— Distribution of Total Number of Progeny.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Renewal Processes — Renewal Process in Discrete and Continuous Time—Renewal Interval— Renewal Function and Renewal Density — Renewal Equation — Renewal theorems: Elementary Renewal Theorem.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Stationary Processes: Application to Time Series. Auto-covariance and Auto-correlation functions and their properties. Moving Average, Autoregressive, Autoregressive Moving Average. Basic ideas of residual analysis, diagnostic checking, forecasting.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] S.M Ross, Stochastic Processes, Wiley India Pvt. Ltd,, 2008.

[2] J. Medhi, Stochastic Process, 2nd Edition, Wiley, 2019.

[3] A.K. Basu, Introduction to Stochastic Process, 2nd Edition, CRC Press, 2017.

[4] B.R. Bhat, Stochastic Models: Analysis and Applications, 2nd Edition, CRC Press, 2018.

ESE - 40%

Students who complete this course will gain a solid foundation in how to apply and understand Bayesian statistics and how to understand Bayesian methods vs frequentist methods.  Topics covered include: an introduction to Bayesian concepts; Bayesian inference for binomial proportions, Poisson means, and normal means; modelling.

Basics of minimaxity - subjective and frequentist probability - Bayesian inference - prior distributions - posterior distributions -  loss function - the principle of minimum expected posterior loss - quadratic and other common loss functions - advantages of being Bayesian - Improper priors - common problems of Bayesian inference - Point estimators - Bayesian confidence intervals, testing – credible intervals         

Practical Assignments:                                                                 

1. Construction of prior, conditional and posterior probabilities for the chosen data set
2. Computation minimum expected posterior loss.
3. Computation of Bayesian confidence intervals.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Two Equivalent Ways of Using Bayes' Theorem - Bayes' Theorem for Binomial with Discrete Prior - Important Consequences of Bayes' Theorem - and Bayes' Theorem for Poisson with Discrete prior.   

Practical Assignments:                                                     

1. Bayes Classification
2. Examples on Binomial distribution with discrete prior.
3. Examples of Poisson distribution with discrete prior.

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Using a Uniform Prior - Using a Beta Prior - Choosing Your Prior - Summarizing the Posterior Distribution - Estimating the Proportion - Bayesian Credible Interval

Comparing Bayesian and Frequentist Inferences for Proportion:

Frequentist Interpretation of Probability and Parameters - Point Estimation - Comparing Estimators for Proportion - Interval Estimation - Hypothesis Testing - Testing a One-Sided Hypothesis - Testing a Two-Sided Hypothesis.                

Bayesian Inference for Poisson: Some Prior Distributions for Poisson - Inference for Poisson Parameter.                       

Practical Assignments:                                                                 

1. Estimation of binomial proportion using uniform prior distribution.
2. Estimation of binomial proportion using beta prior distribution.
3. Estimation of Poisson parameter using some prior distributions

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Bayes' Theorem for Normal Mean with a Discrete Prior - Bayes' Theorem for Normal Mean with a Continuous Prior - Normal Prior, Bayesian Credible Interval for Normal Mean - Predictive Density for Next Observation.

Practical Assignments:     

1.     Bayes estimator for Normal Mean with a Discrete Prior.

2.     Bayes estimator for Normal Mean with a Continuous Prior.

3.     Bayes Credible interval for the normal mean

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Analytic approximation - E-M Algorithm - Monte Carlo sampling - Markov Chain Monte Carlo Methods - Metropolis-Hastings Algorithm - Gibbs sampling: examples and convergence issues.      

Practical Assignments:     

1. E-M algorithm.
2. Monte-Carlo sampling.
3. Gibbs Sampling.
4. Markov Chain Monte-Carlo application

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Jim, A., Bayesian Computation with R, 2nd Edition, Springer, 2009.

[2] Christensen R, Johnson, W., Branscum, A. and Hanson T. E., Bayesian Ideas and Data Analysis: An Introduction for Scientists and Statisticians, Chapman & Hall, 2011.

[3] Congdon, P, Bayesian Statistical Modeling, Wiley, 2006. 

[4] Ghosh, J. K., Delampady M. and T. Samantha, An Introduction to Bayesian Analysis: Theory & Methods, Springer, 2006.

[5] Rao. C.R. and Day. D., Bayesian Thinking, Modeling & Computation, Handbook of Statistics, Vol. 25. Elsevier, 2006.

ESE - 40%

Data ethics and privacy protection are critical components of responsible AI development and deployment. This course provides a comprehensive overview of the ethical considerations and privacy issues that arise in the context of AI strategy. Students will explore the ethical frameworks, principles, and guidelines relevant to data-driven technologies. They will also examine strategies for implementing privacy-preserving techniques and ensuring compliance with relevant regulations. Through case studies and discussions, students will develop a deeper understanding of the ethical implications of AI and strategies for promoting responsible use of data in AI applications.

Overview of data ethics and privacy in AI - Importance of ethical considerations in AI strategy - Key concepts: fairness, accountability, transparency, and explainability

Privacy Protection in AI Systems -  Privacy threats and challenges in AI applications - Privacy-preserving techniques: anonymization, encryption, and differential privacy - Privacy by design principles

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Utilitarianism, deontology, virtue ethics, and consequentialism - Ethical principles and guidelines for AI development - Application of ethical frameworks to AI strategy

Legal and Regulatory Landscape - Overview of data protection laws and regulations (e.g., GDPR, CCPA) - Compliance requirements for AI systems - Impact of regulatory frameworks on AI strategy

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Ethical considerations in data collection and processing - Data governance and responsible data management practices - Risk assessment and mitigation strategies - Best practices for ethical data collection and usage in AI systems - Addressing bias and fairness in data-driven decision-making

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Importance of transparency in AI decision-making - Techniques for explaining AI predictions and decisions - Challenges and limitations of explainable AI -  Methods for ensuring transparency and accountability in AI development and deployment

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Analysis of real-world ethical dilemmas in AI - Ethical decision-making frameworks - Discussion and debate on ethical implications of AI applications

Emerging Issues in Data Ethics and AI - Exploring emerging issues and trends in data ethics and AI - Anticipating future challenges and opportunities

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Bruce Schneier, "Data and Goliath: The Hidden Battles to Collect Your Data and Control Your World", W. W. Norton & Company, 2016.

[3] Cathy O'Neil, "Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy", Crown, 2016.  

[4] Markus Dubber, Frank Pasquale, and Sunit Das, "Oxford Handbook of Ethics of AI”, OUP USA 2021.  

[5] Shoshana Zuboff , "The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power", PublicAffairs, 2019.

[2] Helen Nissenbaum, "Privacy in Context: Technology, Policy, and the Integrity of Social

Life",  Stanford University Press, 2009.

[3] Joanna J. Bryson, "Ethics for Robots: How to Design a Moral Algorithm".  

[4] George Reynolds, "Ethics in Information Technology", Course Technology Inc, 5th edition, 2014. 

[5] Lee A. Bygrave,  "Data Privacy Law: An International Perspective", Oxford University Press,  2014.  

[6] Meredith Broussard, "Artificial Unintelligence: How Computers Misunderstand the World", MIT Press, 2018.

[7] Frank Pasquale, "The Black Box Society: The Secret Algorithms That Control Money and Information", Harvard University Press, 2016.

ESE - 40%

The subject is intended to give the knowledge of Big Data evolving in every real-time applications and how they are manipulated using the emerging technologies. This course breaks down the walls of complexity in processing Big Data by providing a practical approach to developing Java applications on top of the Hadoop platform. It describes the Hadoop architecture and how to work with the Hadoop Distributed File System (HDFS).

Concepts of Data Analytics: Descriptive, Diagnostic, Predictive, Prescriptive analytics - Big Data  characteristics: Volume, Velocity, Variety, Veracity of data - Types of data: Structured,  Unstructured, Semi-Structured, Metadata - Big data sources: Human-Human communication,  Human-Machine Communication, Machine-Machine Communication - Data Ownership - Data  Privacy.  

Standard Big data architecture - Big data application - Hadoop framework - HDFS Design goal - Master Slave architecture - Block System - Read-write Process for data - Installing HDFS - Executing in HDFS:  Reading and writing Local files and Data streams into HDFS - Types of files in HDFS - Strengths and  alternatives of HDFS - Concept of YARN. 

Lab Practice

1.     Exercise on Reading and Writing Local files into HDFS

2.     Exercise on Reading and Writing Data streams into HDFS

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Introduction to MapReduce - Sample MapReduce application: Wordcount - MapReduce Data types and  Formats - Writing MapReduce Programming - Testing MapReduce Programs - MapReduce Job  Execution - Shuffle and Sort - Managing Failures - Progress and Status Updates.

Lab Practice

1.     Exercise on MapReduce applications

2.     Exercise on writing and testing MapReduce Programs

3.     Exercise on Shuffle and Sort

4.     Exercise on Managing Failures

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Hive Architecture - Components - Data Definition - Partitioning - Data Manipulation - Joins, Views and  Indexes - Hive Execution - Pig Architecture - Pig Latin Data Model - Latin Operators - Loading Data - Diagnostic Operators - Group Operators - Pig Joins - Row Level Operators - Pig Built-in function - User defined functions - Pig Scripts. 

Lab Practice

1.     Exercise on Hive Architecture

2.     Exercise on Pig Architecture 

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

Stream processing Models and Tools - Apache Spark - Spark Architecture: Resilient Distributed Datasets,  Directed Acyclic Graph - Spark Ecosystem - Spark for Big Data Processing: MLlib, Spark GraphX,  SparkR, SparkSQL, Spark Streaming - Spark versus Hadoop  

Lab Practice

1.      Exercise on Directed Acyclic Graph

2.      Exercise on Spark using MLlib, Spark GraphX

3.      Exercise on Spark using SparkR, Spark Streaming

Teaching  / learning Strategy: Lecture  / Discussion / Presentation / Problem solving / Class Activity

[2] Chandramouli Subramanian, Asha A George, C R Rene Robin, D Doreen Hephzibah Miriam, J Jasmine Christina Magdalene, ”Big Data Analytics”, Universities Press (India) Private Limited, 2024.

[2] Julián Luengo, Diego García-Gil, Sergio Ramírez-Gallego, Salvador García, Francisco Herrera, Big Data Preprocessing: Enabling Smart Data. Springer Nature Publishing, 2020.

[3] Seema Acharya, Subhasini Chellappan, Big Data and Analytics. 2nd Edition, Wiley India Pvt Ltd, 2019.

ESE - 40%

This course is designed to provide students with the opportunity to integrate and apply the knowledge, skills, and competencies acquired throughout their academic journey. This course serves as the culmination of the student's academic experience, enabling them to tackle real-world problems, explore innovative solutions, and demonstrate their readiness for professional endeavors. Through a blend of research, analysis, critical thinking, and practical application, students will engage in a comprehensive project that reflects their chosen field of study or specialization.

This project is to be taken up either in the industry or in an R&D organization. The students will engage in a comprehensive project that reflects their chosen field of study or specialization.

ESE - 40%