Schedule and Syllabus


This course meets Mondays (8:00am - 9:55am) and Tuesdays (from 12:00 - 12:55pm)
Note: GBC = "Deep Learning", I Goodfellow, Y Bengio and A Courville, 1st Edition Link
Module 01 Lecture 1 Monday
Jan 04		 Introduction, Linear Algebra Primer
Course logistics and overview. Linear Algebra review. 		[slides (pptx)]
Lecture 2 Tuesday
Jan 05		 Linear Algebra Primer (contd.) Vector Calculus Review
Brief review of concepts from Vector Calculus. 		[slides (pptx)]
Lecture 3 Monday
Jan 11		 Optimization
Types of errors, bias-variance trade-off, overfitting-underfitting, brief review of concepts from optimization, variants of gradient descent, momentum based methods. 		[slides (pptx)]
Class test-I on Module 01 on Jan 19 (Tuesday)
Module 02 Lecture 4 Tuesday
Jan 12		 Artificial Neural Networks
Linear and Logistic Regression. Basic concepts of artificial neurons, single and multi layer perceptrons, perceptron learning algorithm, its convergence proof, different activation functions, softmax cross entropy loss function. 		[slides (pdf)]
[slides (pdf)]
[slides (pdf)]
[slides (pdf)]
Lecture 5 Monday
Jan 18
Lecture 6 Monday
Jan 25
Module 03 Lecture 7 Monday
Feb 01		 ConvNets
ConvNets: Basic concepts of Convolutional Neural Networks starting from filetering. Convolution and pooling operation and arithmatics of these. ConvNet Architectures: Discussions on famous convnet architectures - AlexNet, ZFNet, VGG, C3D, GoogLeNet, ResNet, MobileNet-v1, EfficientNet
Discussion on regularization, Dropout, Batchnorm etc. 		[slides (pptx)]
[slides (pdf)]
[slides (pptx)]
Lecture 8 Tuesday
Feb 02
Lecture 9 Monday
Feb 08
Class test-II on Module 02 and first part of Module 03 on Feb 09 (Tuesday)
Module 04 Lecture 10 Monday
Feb 15		 Recurrent Architectures
Discussion on Recurrent Neural Networks (RNNs), Long-Short Term Memory (LSTM) architectures, Attention. 		[slides (zip)]
Lecture 11 Monday
Feb 22
Module 05 Lecture 12 Tuesday
Mar 08		 Applications: Detection and Segmentation
Discussion on detection, segmentation problem definition, challenges, Evaluation, Datasets and Localization by regression. Discussion on detection as classificaion, region proposals, RCNN and YOLO architectures, fully convolutional segmentations, Mask-RCNNs. 		[slides (pdf)]
[slides (pdf)]
Lecture 13 Monday
Mar 15
Class test-III on second part of Module 03 and Module 04 on Mar 09 (Tuesday)
Module 06 Lecture 14 Monday
Mar 16		 NLP Basics and Transformers
word2vec, Transformers and Applications to NLP 		[slides (pdf)]
[slides (pdf)]
Lecture 15 Tuesday
Mar 22
Lecture 16 Monday
Mar 23
Class test-IV on Module 06 on Mar 30 (Tuesday)
Module 07 Lecture 17 Monday
Apr 05		 Explainability and Bias
Discussion on explainability and bias in Deep Learning system. The need for explanation, introspection vs justification, activation maximization and activation map based explanation generation, Black-box explanation generation etc. Bias in AI and in image captioning task.
Lecture 18 Tuesday
Apr 06
Module 08 Lecture 19 Monday
Apr 12		 GAN
Stochastic AE, VAE, GAN
Lecture 20 Tuesday
Apr 13