Project Grant 2234227

The Georgia State University Research Foundation Inc. was awarded a two-year, $255,245 Project Grant from the National Science Foundation Division of Information and Intelligent Systems to develop green granular neural networks with fast, FPGA-based incremental transfer learning algorithms. Under the Computer and Information Science and Engineering federal grant program, the awardee will create a novel shallow software-hardware machine learning system using green and energy efficient field programmable gate array hardware. This is intended to significantly reduce both carbon dioxide emissions and energy consumption from machine learning compared to traditional CPU and GPU methods. Specifically, the awardee will develop a shallow granular neural network tree incorporating new incremental transfer learning algorithms solved directly on an FPGA. The resulting explainable and efficient machine learning system is intended for real-time green computing applications. Undergraduate and graduate students, including those from underrepresented groups, will participate in research to advance intelligent green computing education and help train the next generation of machine learning workforce.