Salim El Rouayheb receives ARL Grant to Design Machine Learning Algorithms for Tactical Edge Communication Networks

ECE Associate Professor Salim El Rouayheb is the recipient of a new award from the Army Research Lab (ARL) for the project titled "Machine Learning Algorithms and Optimizations for Resource-Constrained Tactical Edge."  Dr. El Rouayheb is the Rutgers PI on a three-year $1.2 million collaborative effort with the University of Illinois at Chicago (UIC), with Rutgers share of the award being $497,000.
In this project, Dr. El Rouayheb and his collaborators aim to improve the speed and accuracy of AI-based edge solutions through a  model-distributed framework for faster and lightweight AI/ML inferencing. The focus will be on understanding the potential and impacts of model distributed inferencing for resource-constrained tactical edge networks. The project will first investigate early exit mechanisms based on E2CM, a novel Early Exit technique based on Class Means of samples.  Moreover, the project will study decentralized learning algorithms, for real-time learning and inferencing. The security of these algorithms against passive and adversarial attacks will be investigated using secure coded computation. Finally, the project will develop a heterogeneous distributed computing testbed consisting of commodity end devices and edge servers to implement and validate the obtained algorithms.

Congratulations, Salim!