ED 

Emon Dey

Postdoctoral Appointee at Argonne National Laboratory
9700 S Cass Ave, Lemont, IL 60439

edey [at] anl [dot] gov
Google Scholar   GitHub   Resume   LinkedIn

About

I am working as a Postdoctoral Appointee at the Argonne National Laboratory. I received my PhD. in Computer Information Systems from the University of Maryland, Baltimore County (UMBC). I completed my tenure as a Research Assistant at the Center for Real-time Distributed Sensing and Autonomy (CARDS). I was also affiliated with Mobile, Pervasive, and Sensor Computing (MPSC) Lab and advised by Dr. Nirmalya Roy. My research interests mainly focus on the area of Artificial Intelligence and Robotics, particularly on developing deep model compression algorithms, robust communication protocols for Federated Learning, and building packages on Robot Operating Systems (ROS).

I received my M.S. degree in Information Systems from UMBC in 2022. Before that, I received my Bachelor's degree in Electrical and Electronic Engineering from Chittagong University of Engineering & Technology. My thesis Supervisor was Sharmin Majumder and I have worked on developing a novel Image Watermarking algorithm.

News

Current Research

My research focuses on designing and optimizing scalable training architectures for large-scale Vision Transformer models on High-Performance Computing (HPC) infrastructures, with applications spanning coherent X-ray imaging and smart power grid analytics. Also, I am developing an intelligent scheduling algorithm to improve the communication efficiency of Asynchronous FL. During my PhD. I have developed a viable algorithm to make the data communication protocol robust for Federated Learning (FL) scenario under a contested environment. To be specific, the proposed algorithm is useful in reducing the packet loss due to the higher velocity difference among agents participating in building the FL model.

Education

Publications

Under Review

Under Preparation

Published Works

Skills

Research Experience

Teaching Experience

Awards

Media Coverage

Professional Activities

Helpful Resources

Edge Computing

Federated Learning and Model Compression

Robot Operating Systems (ROS)