Overview of topics to be covered in the Programme 

Motivation for the Programme 

Overview of the Programme 

Expected Outcomes of the Programme 

Brief about software/tools

Linear algebra: Vectors, matrices, inner products, matrix-vector multiplication, eigen values/vectors, singular value decomposition 

Calculus: Differentiation (single/multiple variables, vectors, and matrices), unconstrained and constrained optimisation (Lagrangian multiplier) 

Probability Theory: Discrete and continuous random variables, probability distributions, Bayes' rule, Gaussian density function, conditional probability 

Statistics: Descriptive and inferential statistics, hypothesis testing, probability distributions

Python: Pre-read 

Python details: Python syntax, factors, NumPy, Scipy, Pandas, Data Visualization, Scikit Learn, Pytorch,Matplolib, Seaborn Tensorflow, Deployment and productionisation 

Advanced python techniques: generators, iterators, decorators, context managers, performance optimisation techniques. Demo on Python tools, python packages, pytorch, scikit learn, tensorflow, demo of deployment on python, demo on advanced python techniques

EDA: Data types and variables, central tendency and dispersion 

Five-point summary and skewness, Box-plot, covariance and correlation, encoding, scaling and normalisation.  

Focus on pre-processing, missing values, working with outliers, demo on EDA

NLP and text processing applications: Text classification, parts-of-speech tagging, named entity recognition, text summarization, text question answering, machine translation. Demo on sentiment analysis, chatbot creation and text-to-text translation 

Image and video processing applications: Image classification, image annotation, image captioning, video classification, video captioning, visual question answering, visual common-sense reasoning 

Speech processing applications: Speech recognition, speaker recognition, speech emotion recognition, spoken language recognition, text-to-speech synthesis, speech-to-speech translation 

Supervised learning 

Unsupervised learning 

Semi-supervised learning 

Active learning 

Self-supervised learning 

Transfer learning 

Domain adaptation, Zero-shot 

One-shot and Few-shot learning; Federated learning

Linear model for regression 

Supervised learning 

Parameter estimation 

Overfitting 

Regularisation 

Ridge regression 

K-nearest neighbour classifier 

Bayes classifier 

Normal density function 

Decision surfaces 

Naïve Bayes classifier 

Maximum likelihood estimation 

Gaussian mixture model 

Distance of a point to a hyperplane 

Margin of a separating hyperplane 

Hard-margin SVM 

Soft-margin SVM 

Kernel functions 

Multi-class classification using SVMs

Principal component analysis 

Fisher discriminant analysis

Construction of decision tree for classification 

Random forest classifier 

Bagging 

Boosting 

AdaBoost 

Applications of Ensemble methods 

K- -Means clustering 

Hierarchical clustering 

Applications of Clustering Techniques

McCulloch-Pitts neuron 

Perceptron learning rule 

Sigmoidal activation function 

ReLU activation function 

Softmax activation function 

Multilayer feedforward neural network 

Error backpropagation method 

Gradient descent method 

Stochastic gradient descent method 

Stopping criteria, Logistic regression-based classifier 

Focus on Deep Learning using Tensorflow and Keras, understanding Feedforward neural network, back propagation, gradient descent and logistic regression

Generalized delta rule 

AdaM based optimizer 

Regularization: Drop-out, Drop-connect, Batch normalization 

Basic CNN architecture, Rectilinear Unit (ReLU), 2-D Deep CNNs: LeNet, VGGNet, GoogLeNet, ResNet 

Image classification using 2-D CNNs 

3-D CNN for video classification 

1-D CNN for text and audio processing 

Object localization and detection algorithms – YOLO, Image Segmentation, and UNet 

Architecture of an RNN, Unfolding an RNN, Backpropagation through time 

Long short-term memory (LSTM) units 

Gated recurrent units 

Bidirectional RNNs 

Deep RNNs 

Structure of GAN, types of GAN models, applications of GAN models

Attention mechanism 

Transformer architecture 

BERT (Bidirectional Encoder Representations from Transformers) 

ViLBERT 

GPT (Generative Pre-trained Transformer) 

Applications of transformer models

Applications of Gen AI in different domains 

Examples of prompt engineering, fine tuning and API creation and integration 

Markov Decision Processes (MDPs) 

Q-Learning and Deep Q Networks (DQN) 

Actor-Critic models 

Exploration vs. Exploitation strategies 

The capstone project is a comprehensive, real-world assignment in which participants apply their knowledge and skills to solve industry-specific problems  

It integrates concepts from their coursework, encouraging critical thinking and innovation  

Capstone projects help participants gain hands-on experience, making them industry-ready by demonstrating their ability to tackle complex challenges in a professional setting 

Ethical considerations (banking, ecommerce sectors); pushing code to repository 

Responsible AI 

Explainable AI 

Registry, Model & Data Monitoring

Understanding cloud infrastructure essentials 

Cloud-based ML Services and Databases 

Containerization 

Cloud enablement - scalability and flexibility 

Understanding emerging themes: FaaS, Edge Computing 

Federated Learning 

AutoML 

Explainable AI 

Cloud ML-Ops 

Deployment on Gemma models on Vertex AI and Kubernetes engine 

Scaling with AWS