Natural Language Processing (Jan 2022)

This contains the details of the tutorial sessions (2 hours each) conducted during this course. Besides these tutorials, content was made for the class in form of ppts, and assignments, and evaluating them.

Tutorial 1: Mathematics for Machine Learning 1 (Linear Algebra): Vector Spaces, Fundamental Subspaces, Projections, Inner products, Norms, etc.

Tutorial 2: Mathematics for Machine Learning 2 (Probability and Statistics): Random Variables, pdfs and cdfs, distributions: Bernoulli, Binomial, Exponential, Normal (Univariate and Multivariate cases), Transformation of Random Variables

Tutorial 3: Information Theory Ideas, Conventional features in NLP: Entropy, Classical models (basics) like WFSTs, Bag of Words, TF-IDF features

Tutorial 4: Introduction to Deep Learning: MLPs, basics of multivariate calculus, computation graphs, backpropagation

Tutorial 5: Sequential Models 1: problems with MLPs in ANNs, introduction to RNNs, backpropagation in time, vanishing gradients, GRUs

Tutorial 6: Sequential Models 2: LSTMs, BLSTMs, Using Sequential models as a language model, introduction to word embeddings

Tutorial 7: Word Embeddings: problems with one-hot encoding, demonstration of similarity with tSNE, Embedding layer, How to learn them, self supervised embedding learning: Word2Vec, GloVE

Tutorial 8: Transformers (basics): Problems with LSTMs in translation of long sentences, introduction to self attention, key-query framework, positional embeddings, using them in transformers like BERT.

You can get the tutorial materials here. Certificate