EE 5611: Machine Learning Applications for Wireless Communications

Course Syllabus

The course contents and the grading procedure is available here. The following grading policy is followed for this course.

  • Project implementation - 40%

  • In-class presentation - 40%

  • Kaggle Assignments - 10%

  • Class participation - 10%


The course is organized in two parts

  • Overview of ML and its applications to various wireless problems.

  • Study of recent papers in the wireless communications domain that explored machine learning and deep learning concepts.

Part 1

The following is the set of lectures in the first part of the course.

  1. Introduction to the course

    1. What is machine learning, what is communication systems, how can these two fields be used to address relevant problems?

  2. Introduction to Machine Learning

    1. Overview of supervised, un-supervised, reinforcement learning

  3. Introduction to Communication Systems

    1. Various aspects of communication systems, wireless system design, where machine learning can be applicable in various OSI layers of a communication system, how real time schedulers can benefit from advanced machine learning techniques

  4. Connections between signal processing, adaptive filtering and machine learning

    1. Invited talk, Dr. K Sri Rama Murty, HoD, EE IITH

    2. connections between weiner filtering and regression techniques in ML, filtering techniques such as FIR, IIR and the deep neural netwpork architectures such as CNN, RNN

  5. Supervised Learning and its applications in wireless systems

    1. Applications in modulation classification, adaptive modulation and coding mechanisms for wireless systems, etc.

  6. Evolution in AI Thinking: Building Products of the Next Decade

    1. Invited Talk, Mr. Raghuram Lanka, AVP, Jio Labs

  7. Un-supervised Learning and its applications in wireless systems

    1. Use of principal component analysis in massive mimo system design, auto encoders in wireless communication transceiver design, etc.

  8. Connections between wireless communications and machine learning

    1. Invited talk, Dr. Sumohana S. Channappayya, EE, IITH

    2. Connections between hidden markov model (commonly used for NLP) and viterbi algorithm commonly used in convolutional decoders in wireless systems

Part 2

The next set of lectures are the student presentations where we studied recent papers from the field of wireless communications where ML and deep learning tools were innovatively used for addressing a wide variety of problems.

  1. An Introduction to deep learning for the physical layer

  2. Deep Architectures for Modulation Recognition

  3. On Deep Learning-Based Channel Decoding

  4. An Improved WiFi Indoor Localization Method Combining Channel State Information and Received Signal Strength

  5. Online Learning in Bittorrent System

  6. End-to-End Learning of Communications Systems Without a Channel Model

  7. Channel State Information Prediction for 5G Wireless Communications: A Deep Learning Approach

  8. Predicting Wireless Channel Features using Neural Networks

  9. Deep Learning Based MIMO Communications

  10. Deep Reinforcement Learning for DynamicMultichannel Access in wireless networks

  11. Deep Reinforcement Learning Autoencoder with Noisy Feedback

Kaggle Assignments

  1. Feature-based Modulation Classification in AWGN Noise

  2. Deep Learning-based Modulation Classification in AWGN Noise

  3. Deep Learning-based Modulation Classification in Rayleigh Multipath Channels

  4. Deep Learning-based Modulation Classification in IIR-type Channels


  • Probability and Statistics

  • Linear Algebra

  • Machine Learning

  • Communication Systems

  • Wireless Communications

Online Resources

Machine Learning For Communications Emerging Technologies Initiative
Best Readings in Machine Learning in Communications
Machine Learning, Coursera