Motivation for Programme, Overview of Programme, Expected Outcomes of Programme

Function approximation (Regression), Classification, Clustering, Ranking, Information retrieval Text processing applications - Text classification, Parts-of-speech tagging, Named entity recognition, Text summarization, Text question answering, Machine 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;            
Data representation - Feature extraction, Representation learning, Embeddings

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

Review of basics of mathematical topics Linear Algebra: Vectors and Matrices, Inner product of vectors, Matric multiplication by a vector, Eigen values and vectors of a matrix, Singular value decomposition of a matrix.
Calculus: Differentiation with one variable, Differentiation with multiple variables, Differentiation of a vector and a matrix, Unconstrained optimization problem solving.
Probability and Statistics: Random variables, First order and second order statistics, independent variables, Uncorrelated variables, Sum and Product Laws of probability, Probability distributions, Bayes rule.

Linear model for regression, Supervised learning, Parameter estimation - Maximum likelihood method, Overfitting, Regularization, Ridge regression, Lasso

K-nearest neighbours’ classifier, Bayes classifier, Normal density function, Maximum likelihood estimation, Gaussian mixture model, Naïve Bayes classifier, Decision surfaces, Dimension reduction methods - Principal component analysis, Linear discriminant analysis

McCulloch-Pitts neuron, Perceptron learning rule, Perceptron convergence theorem, Sigmoidal neuron, Softmax function, Multilayer feedforward neural network, Error backpropagation method, Gradient descent method, Stochastic gradient descent method, Stopping criteria, Logistic regression based classifier

Deep feedforward neural networks (DFNNs), Optimization methods: Generalized delta rule, AdaGrad, RMSProp, Adadelta, AdaM, Second order methods;  Regularization methods: Dropout, Dropconnect; Batch normalization

Basic CNN architecture, Rectilinear Unit (ReLU), 2-D Deep CNNs: LeNet, AlexNet, VGGNet, GoogLeNet, ResNet; Image classification using 2-D CNNs; 3-D CNN for video classification; 1-D CNN for text and audio processing; Vector of Linearly Aggregated Descriptors (VLAD) method for aggregation - NetVLAD

Architecture of an RNN, Unfolding an RNN, Backpropagation through time, Vanishing and exploding gradient problems in RNNs, Long short term memory (LSTM) units, Gated recurrent units, Bidirectional RNNs, Deep RNNs

Encoder-decoder paradigm, Image and video captioning models, Machine translation, Text processing models, Representation of words: Word2Vec, GloVe

Attention based models, Scale dot product attention, Multi-head attention (MHA), Self-attention MHA,  Cross-attention MHA, Position encoding, Encoder Decoder module in a transformer, Sequence to sequence mapping using transformer, Machine translation using transformer model, Vision transformer for image classification, Video captioning using transformer model, Bidirectional encoder representations from transformers (BERT) model for text processing, Pre-training a BERT model, Fine tuning a BERT model for text processing tasks, Vision-and-Language BERT (ViLBERT) for image and video processing tasks, Text and Visual question answering and reasoning using transformer models

Image generation models, Architecture and training of a GAN, Deep convolutional GAN, Cyclic GAN, Conditional GAN, Super-resolution GAN, Applications of GANs for image processing

Introduction to reinforcement learning, Markov decision processes, Policy gradients, Temporal difference learning, Q-learning, Deep reinforcement learning - Deep policy gradient, Deep Q learning; Text processing using deep reinforcement learning - Text classification, Text summarization