電子情報通信学会ソサイエティ大会講演要旨
C-8-13
Design and Evaluation of Deep Learning Accelerators Using Superconductor Logic Families
◎Qiuyun Xu(Yokohama National Univ.)・Yanzhi Wang(Northeastern Univ.)・Naoki Takeuchi・Nobuyuki Yoshikawa(Yokohama National Univ.)
In this study, we present the first design of deep learning accelerator using adiabatic superconductor technologies. Evaluation indicates that it can achieve around 88% accuracy in the MNIST data set for handwritten digit recognitions.