
Dr. Yinan Wang’s proposal, “Model Optimization and Hardware-aware Neural Architecture Search for Spatiotemporal Data Mining” was selected as a Future of Computing Research Collaboration (FCRC) project, a partnership between RPI and IBM. This research program is focused on developing a comprehensive framework for hardware-aware neural architecture search and model optimization for on-edge spatiotemporal data mining. More specifically, (1) spatiotemporal sparsity is introduced to develop a novel lightweight neural network layer, and (2) the hardware architecture and neural networks are co-designed for improved efficiency, accuracy, and energy consumption. Dr. Liu Liu from the ECSE Department will collaborate on this project.