Empower the creation, training and use of more sophisticated Deep Learning through higher performance and higher energy efficient hardware units. Design hardware units that deliver the best by far performance and better energy efficiency over existing designs targeting a broad spectrum of devices from mobile to data center server units. Our approach already delivers 2x over the best design available today. Our approach is different than everything else that is available in that we also exploit properties of the content being calculated upon open up opportunities that no other design currently exploits.

Team Members

Andreas Moshovos
Professor, Electrical & Computer Engineering

Jorge Alberico
Postdoctoral Fellow, Electrical & Computer Engineering

Patrick Judd
PhD Candidate, Electrical & Computer Engineering