Energy-efficient AI

We aim at bringing AI from data centers to the edge. Lower energy consumption will improve battery-powered devices autonomy. Moreover, no need for data transfer over the network will ultimately enhance privacy and provide better data security.

Publications

. Stochastic Computing for Hardware Implementation of Binarized Neural Networks. IEEE Access, 2019.

Preprint PDF Project

Talks

Medical Applications of Low Precision Neuromorphic Systems
Sep 26, 2019