The National Science Foundation (NSF) awarded a $134,992 Project Grant (CFDA 47.070 - Computer and Information Science and Engineering) to Rensselaer Polytechnic Institute (RPI) to develop methods for reducing energy consumption in large artificial intelligence (AI) systems. The key objectives of this 5-year project are to: 1) leverage dynamic connectivity to reduce redundancy in AI models by adapting them to specific tasks and data; 2) provide architectural support for elastic processing through heterogeneous computing architectures; and 3) enable hardware-aware model adaptation to improve efficiency. This project aims to overcome the energy challenges associated with increasingly large AI models and computing resources by reducing computational costs while maintaining model performance, advancing future paradigms for energy-efficient and scalable AI systems. RPI plans to integrate the research activities with educational initiatives to engage students and broaden participation among local communities and underrepresented groups.
Generated 3/4/25, 6:25 AM