E0 270-O Machine Learning 3:1 (January 2023)

Course Instructor: Ambedkar Dukkipati, CSA

Course description: The aim of the course is to provide a unified view of AI (Artificial Intelligence) and machine learning methods. This course stresses on foundations and covers important topics in three elements of machine learning: supervised learning, unsupervised learning, and reinforcement learning. Further this course provides basics of deep learning and emphasizes how deep neural networks play key role in various problems as the powerful function approximators. Assignments will involve programming and requires knowledges of python.
Webpage will be made available here

Syllabus

Supervised Learning: Classification with Bayes rule, learning as optimization, linear regression, logistic regression, probabilistic view: ML and MAP estimates, gradient descent, hyperplane-based classifiers, perceptron, kernel methods, feedforward neural networks, backpropagation algorithm, CNN (Convolutional Neural Network), RNN and LSTMs.
Unsupervised Learning: Principal component analysis, clustering methods, undirected graphical models, MCMC (Markov Chain Monte Carlo) and Gibbs sampling, latent variable and mixture models, EM (Expectation-Maximisation) algorithm, Deep generative models.
Reinforcement Learning: Introduction to sequential decision making and online learning, Markov decision processes and dynamic programming.

Textbooks / References

1. C. M. Bishop: Pattern Recognition and Machine Learning. 2006
2. Hastite T, Tibshirani R and Friedman J: The Elements of Statistical Learning: Data Mining, Inference and Prediction, 2009
3. Haykin. S: Neural Networks and Learning Systems, 2009.
4. Goodfellow, Bengio, Courville: Deep Learning, 2017

Prerequisites: There are no formal prerequisites.

Grading:

• Homeworks 40%
• Midterm 30%
• Final exam 30%.