Abolfazl Hashemi

Assistant Professor of ECE

Machine Intelligence and Networked Data Science Group (MINDS)

The Institute for Emerging Core Methods in Data Science (EnCORE)

The Center for Innovation in Control, Optimization, and Networks (ICON)

The Center for Education and Research in Information Assurance and Security (CERIAS)

I am an Assistant Professor at The Elmore Family School of Electrical and Computer Engineering at Purdue University and the director of MINDS Group. Previously, I was a Postdoctoral scholar at the Oden Institute for Computational Engineering and Sciences at the University of Texas at Austin. I received my PhD and MSE degrees in Electrical and Computer Engineering department at UT Austin in 2020 and 2016, and my BS degree in Electrical Engineering from Sharif University of Technology in 2014.

MINDS Group has an opening for a postdoctoral postion on optimization for distributed machine learning. Please see here for more details.

My graduate course on Optimization for Deep Learning will be offered again in Fall 2023.

Research Interests

My research interests include the study of Machine Learning, Online Learning, Learning in Games, and Networked/Distributed Decision-Making from the perspective of optimization. In doing so, we design efficient algorithms with mathematical guarantees to render practical deployment of learning-based systems possible under a variety of considerations such as limited resources, and adversarial attacks (see the Research tab for more details and the Publication tab for a complete list of my papers).

Recent News

September 2023: Our paper Global Update Tracking: A Decentralized Learning Algorithm for Heterogeneous Data is accepted to NeurIPS 2023.
August 2023: Our two papers on High probability Results for Federated Learning and Robust Sensor Selection are accepted to 2023 Allerton Conference on Communication, Control, and Computing
July 2023: Our three papers on High probability Results for Sensor Selection, Improved Results for Noisy Federated Learning, and Online Reinforcement Learning are accepted to 2023 Asilomar Conference on Signals, Systems, and Computers
June 2023: Our journal paper “No-Regret Learning in Dynamic Stackelberg Games” is accepted to IEEE Transactions on Automatic Control.
June 2023: Our journal paper “On the Convergence of Decentralized Federated Learning Under Imperfect Information Sharing” is accepted to IEEE Control Systems Letters.

For the full list of news please check the News tab.

Selected Publications

Global Update Tracking: A Decentralized Learning Algorithm for Heterogeneous Data, The Conference on Neural Information Processing Systems (NeurIPS), 2023.
Faster Non-Convex Federated Learning via Global and Local Momentum, The 2022 Conference on Uncertainty in Artificial Intelligence (UAI), 2022.
Robust Training in High Dimensions via Block Coordinate Geometric Median Descent, The 25th International Conference on Artificial Intelligence and Statistics (AISTATS), 2022.
No-Regret Learning with High-Probability in Adversarial Markov Decision Processes, Conference on Uncertainty in Artificial Intelligence (UAI), 2021.
Submodular Observation Selection and Information Gathering for Quadratic Models, International Conference on Machine Learning (ICML), 2019.

For the full list of publications please check the Publication tab.