Project Grant 2617391

Federal Project Grant Award Summary

The University of Michigan received a $400,000 Project Grant award from the National Science Foundation's Division of Information and Intelligent Systems under the Computer and Information Science and Engineering program (CFDA 47.070), effective September 1, 2026, through August 31, 2029. This collaborative research initiative addresses fundamental challenges in enabling Spiking Neural Networks (SNNs) to efficiently handle long-range dependencies and generative tasks for edge-based artificial intelligence systems. The research reimagines SNNs as underlying state space models to reduce the computational resource requirements and power consumption of foundation models, particularly for resource-constrained environments such as unmanned aerial vehicles, robots, wearables, and mobile devices where on-chip inference is required due to latency, bandwidth, and privacy constraints.

The project delivers algorithmic innovations and cross-layer research spanning device design, circuit development, systems exploration, and machine learning. Key outputs include neuro-artificial intelligence models that provide orders of magnitude improvements in power and energy efficiency while offering additional benefits such as adversarial robustness. The interdisciplinary research agenda incorporates insights from devices, circuits, machine learning, and dynamical systems, with reciprocal information flow between multiple thrust areas. This initiative is positioned to generate significant impacts for the semiconductor and artificial intelligence industries by enabling efficient implementation of data-intensive machine learning workloads in resource-constrained environments, while serving as a training platform for the next generation of researchers and engineers in neuromorphic computing.