IBM has developed a prototype “brain-like” chip that could make artificial intelligence (AI) more energy efficient.
The chip, which is still in development, uses components that work in a similar way to connections in the human brain. This makes the chip more energy efficient than traditional computer chips, which could lead to longer battery life for smartphones and other devices that use AI.
IBM says the chip could also be used to reduce energy costs in data centers, which are the large warehouses of computers that power AI systems.
“The human brain is able to achieve remarkable performance while consuming little power,” said Thanos Vasilopoulos, a scientist at IBM’s research lab in Zurich, Switzerland. “Our chip is inspired by the brain’s energy efficiency, and we believe it could lead to a new generation of AI chips that are more powerful and more sustainable.”
The chip uses components called memristors, which are analogue devices that can store a range of numbers. This is in contrast to traditional computer chips, which are digital and can only store 0s and 1s.
The use of memristors makes the chip more energy efficient because it allows the chip to process information in a more efficient way. For example, the chip can use less power to store information and to make decisions.
Professor Ferrante Neri, from the University of Surrey, says the development of the brain-like chip is a “significant breakthrough”. “This is a major step towards developing more energy-efficient AI systems,” he said. “The potential benefits are huge, from reducing the environmental impact of AI to making it more affordable for everyone to use.”
Professor Neri cautions that there are still challenges to overcome before the chip can be widely adopted.
The sources for this piece include an article in BBC.