|SALARIS MIRKO||Cycle: XXXVI |
Section: Computer Science and EngineeringMajor Research topic
:A hardware-design framework for efficient neural computations Abstract:
Neural networks are becoming more and more pervasive, but their requirements in terms of computational power and energy consumption hinder their widespread adoption in mobile devices, embedded devices, or where extremely high speed is required. This research project aims to extend their application domains developing a hardware design framework to help developers design both general (i.e., for broad categories) and ad-hoc (i.e., specific for the model) hardware implementations.