Syllabus

Linear Algebra Refresher, Probability Theory Refresher, Digital Signal Processing Refresher

Machine Learning basics: Supervised and Unsupervised learning, Generative and Discriminative models, Classification and Regression (linear models), Evaluation metrics.

Feature selection, dimensionality reduction using PCA; Bayesian classification, Discriminative classifiers: Perceptrons, Multi-layer perceptron, Support Vector Machines; EM Algorithm; K-Means clustering, Classification performance analysis;

Neural Networks and Deep Learning: Multi-class classification and Multi-label classification, Different kinds of non-linearities, objective functions and learning methods

ML for Audio Classification: Probability Models and Expectation Maximization Algorithm, Gaussian Mixture Models, Time Series Analysis: LSTMs and CNNs; ML for Speech Recognition; ML for Image Processing: Transfer Learning, Attention models, Attribute-based learning.

Introduction to Communication Systems: Various aspects of communication systems, wireless system design; ML for Communication: ML applications in wireless systems, Applications in modulation classification, adaptive modulation and coding mechanisms for wireless systems, Use of principal component analysis in massive MIMO system design, auto encoders in wireless communication transceiver design.

This course aims at introducing the students to the fundamentals of machine learning (ML) techniques useful for various signal processing applications and communication systems. It will discuss various mathematical methods involved in ML, thereby enabling the students to design their own models and optimize them efficiently. Prior exposure to ML is not required. The course will be focused on applications in signal processing and communication, and the theory will be tailored towards that end.

Be familiar with foundation of machine learning, and its applications in signal and communication systems.

Understand and analyze various feature selection and classification techniques using machine learning

Implement the machine learning and deep learning methods for multi-class and multi-label classification

Identify, demonstrate and apply the knowledge in signal processing and communication systems

C.M. Bishop, Pattern Recognition and Machine Learning, Springer , 2011

I. Goodfellow, Y, Bengio, A. Courville, Deep Learning, MIT Press , 2016

D. Yu and L. Deng, , Automatic Speech Recognition: A Deep Learning Approach, Springer , 2016

Online Resources:
https://www.comsoc.org/publications/best-readings/machine-learning-communications
https://www.coursera.org/learn/machine-learning
articles and research papers in ML, signal processing and communication conferences (e.g., ICASSP, NeurIPS, ICML, Interspeech, ISMIR, etc.) and journals (e.g., IEEE TASLP, JMLR, IEEE PAMI, IEEE Wireless Comm., etc.).