The Bradley Department of Electrical and Computer Engineering is pleased to announce the winners of the 2021 Bill and LaRue Blackwell Graduate Research Paper Award.

In May 1988, former ECE Department Head Dr. William A. Blackwell donated the revenue to create this award with the purpose of recognizing ECE graduate students who have written the best research paper as judged by the department's faculty.

Aidin Ferdowsi won the first prize award for his dissertation titled "Distributed Machine Learning for Autonomous and Secure Cyber-physical Systems." Ferdowsis dissertation provides a comprehensive set of results that showcase how interdependence among critical infrastructure (smart power system, water systems, gas systems, wireless systems, and transportation systems) can introduce novel security vulnerabilities. Once those vulnerabilities were identified, Ferdowsi provided several solutions, based on game theory and network control, to thwart potential attacks that can leverage those vulnerabilities.

Ferdowsi's advisor Walid Saad, Professor of Electrical and Computer Engineering, said "This work preceded much of the research in the area of edge learning that has now become one of the most coveted research fields. The impact of this paper is seen by the fact that it has already garnered 57 citations although it was published only about a year ago."

Ferdowsi was also awarded the 2021 Outstanding Dissertation - STEM award from the College of Engineering.

Hongyu An won one of two second prize awards for his dissertation titled, "Powering Next-Generation Artificial Intelligence by Designing Three-dimensional High-Performance Neuromorphic Computing System with Memristors." Ans research investigated a self-learning method, named associative memory, to enable the neuromorphic system to potentially learn from surroundings rather than datasets. In his dissertation, associative memory learning is achieved by associating the artificial neural networks (ANNs) together. In this way, the information carried and preprocessed by these ANNs can be correlated. The simulation results demonstrate that the neuromorphic system associates the pronunciation and image of digits together. Moreover, since associative memory is a self-learning method in biological livings, the significance of rebuilding associative memory is not only to reveal a way of designing a brain-like self-learning neuromorphic system but also to explore a method of comprehending the learning/memory mechanism of a nervous system.

An's advisor Cindy Yi, Associate Professor of Electrical and Computer Engineering, stated, "The approaches proposed could address the key challenges in the artificial intelligence and neuromorphic computing, and have significant impacts in the field of semiconductors, integrated circuits, and biological science."

Chiranjib Saha was presented with the final second prize award for his dissertation titled, "Advances in Stochastic Geometry for Cellular Networks." Sahas research developed tools to capture this correlation using ideas from Poisson cluster processes. The general idea is to model the user locations as a cluster process (very similar to how the hotspot users are modeled in simulations by the wireless standardization bodies, such as 3GPP) and then place small cells in those clusters. This naturally couples the locations of the users and base stations. He then developed analytical tools to derive downlink signal to interference and noise ratio (SINR) distribution and other performance metrics for this new setup.

Saha's advisor Harpreet Dhillon, Associate Professor of Electrical and Computer Engineering, commented, "Chiranjib made a remarkable contribution at the intersection of machine learning and stochastic geometry when he demonstrated that a determinantal point process (characterized by a kernel matrix) from stochastic geometry can be used to learn spatial patterns that occur in wireless networks.

Dhillon continued, "The main technical novelty was to come up with an efficient way to learn the DPP kernel matrix from wireless deployment data so that a wireless network deployment can then be treated as a (probabilistic) realization of the 'fitted' DPP."

ECE is proud of the contributions made by its students in advancing each discipline within electrical and computer engineering and is honored to continue the legacies of Bill and LaRue Blackwell.

Written by Greg Atkins