Federal Project Grant Award Summary

The National Science Foundation's Division of Information and Intelligent Systems awarded a $800,000 Project Grant (CFDA 47.070: Computer and Information Science and Engineering) to the Regents of the University of Michigan, effective September 1, 2026, through August 31, 2030. The project develops a neuro-symbolic reasoning framework for verified hardware-software co-design of specialized computing accelerators. The framework integrates artificial intelligence pattern-finding methods with mathematical proof techniques to enable designers to explore significantly larger design spaces while maintaining correctness guarantees essential for safety-critical applications in automotive and medical devices. This approach addresses the growing computational demands of artificial intelligence services and energy-efficiency concerns in semiconductor design.

The deliverables include development of an open-source software tool that reduces design barriers for semiconductor engineers outside large corporations, thereby enhancing domestic semiconductor competitiveness. The research formulates accelerator design as joint synthesis of hardware specializations and software transformations, moving beyond traditional sequential design approaches. Additionally, the project supports workforce development by training graduate and undergraduate students in semiconductor design skills aligned with domestic industry needs. The University of Michigan will conduct this research at its Ann Arbor campus through its Office of Research and Sponsored Projects Division.