Pierre-Louis Sixdenier, Mark Deutel and Prof. Dr.-Ing. Jürgen Teich received the Best Paper Award at this year’s 18th IEEE International Symposium on Embedded Multicore/Many-Core Systems-On-Chip (MCSOC 2025). The award was given for their paper “Early-Exit Neural Architecture Search for Energy-Harvesting Edge Computing.” The award was presented during the event (December 15-18, 2025) in Singapore. In their award-winning work, the authors present a simulation-based framework for searching for neural architectures for exploring EE-DNN architectures and an online energy manager that dynamically decides which early exit to choose based on the available energy and the reliability of the EE-NN predictions, while ensuring energy-neutral operation (ENO).
