IBM's prototype 'brain-like' chip promises greener AI

economictimes.indiatimes.com:

Tech major IBM has said that its prototype 'brain-like' chip could make artificial intelligence (AI) more energy efficient. There are growing concerns about the emissions generated by the large warehouses filled with computers that power AI systems, reports the BBC. According to IBM, its prototype could lead to more efficient, less battery-draining AI chips for smartphones.

The prototype's effectiveness is due to components that function similarly to connections in human brains, the tech major said. Compared to traditional computers, «the human brain is able to achieve remarkable performance while consuming little power», said scientist Thanos Vasilopoulos, based at IBM's research lab in Zurich, Switzerland. Vasilopoulos said that superior energy efficiency would mean «large and more complex workloads could be executed in low power or battery-constrained environments», such as cars, mobile phones and cameras.

«Additionally, cloud providers will be able to use these chips to reduce energy costs and their carbon footprint,» he added. The majority of chips are digital which means that information is stored as 0s and 1s, but the new chip makes use of memristors, which are analogue components that can store a variety of numbers. According to Prof Ferrante Neri of the University of Surrey, memristors are a type of computing that is «nature-inspired» and imitates brain function.