The Best Paper Award at the 21st ACM International Conference on Computing Frontiers in Italy was awarded to researchers from the Electrical and Computer Engineering department. The team, consisting of PhD student Ilknur Mustafazade and professors Naga Kandasamy, PhD, and Anup Das, PhD, presented innovative work on brain-inspired computing systems.
Their paper introduces a new approach to enhance the efficiency of neuromorphic hardware, which mimics the brain’s neural structure for energy-efficient machine learning. The challenge addressed is optimally mapping complex models onto this hardware while minimizing resource waste and inter-system communication.
Unlike previous methods, their technique considers hardware architecture when breaking down models, ensuring effective mapping of subnetworks. The team developed algorithms for two hardware configurations and tested them using various machine learning models.
Results show significant improvement in hardware utilization compared to existing techniques. The paper also explores how model structure affects solution quality and suggests strategies for further enhancement.
This award underscores Drexel’s groundbreaking research in brain-inspired computing. As demand for energy-efficient AI grows, such innovations will be crucial in advancing computing technology.
