This $588,801 Project Grant award from the National Science Foundation's (NSF) Computer and Information Science and Engineering (CISE) program (CFDA 47.070) supports the development of a novel in-memory analog computing (IMAC) architecture and associated design framework to improve the energy efficiency of machine learning (ML) systems.
The key products and services to be delivered under this 3-year award include: (1) developing an IMAC architecture that can perform both matrix multiplication and nonlinear vector operations in the analog domain; (2) designing a hierarchical analog network-on-chip to support large-scale ML workloads on the IMAC architecture; (3) enabling heterogeneous integration of the IMAC technology with existing ML hardware platforms; and (4) creating a fast and accurate simulation framework for the IMAC architecture. The research team will evaluate the scalability, performance, energy efficiency, and accuracy of the developed heterogeneous ML system using standard benchmark suites. This work aims to provide a practical alternative to current high-energy ML systems and support national priorities in artificial intelligence, computing, and nanotechnology.
Generated 2/25/25, 5:09 AM