The University of Southampton in the U.K. is advancing memristors and charge-based processing to make analog and digital technologies adaptable in the nanoscale era.
"If we are to truly compute at the limits of energy efficiency...we need to move towards analog computation techniques whilst being much savvier about how to mix analog and digital signals for maximum effect," contends Southampton's Alexantrou Serb.
Southampton researchers previously demonstrated a single memristor's ability to store 128 different analog levels. Since then, they have added two memristors into the output hierarchy of a classic complementary metal-oxide-semiconductor converter, introducing a potentiometer.
"A direct impact of this research on modern technologies could be the creation of ultra-efficient AI (artificial intelligence) hardware," say the researchers. "AI by nature lends itself to analog implementation of computation much more readily than to the current digital-based techniques."
From Electronics Weekly (UK)
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found