Scientists from Cambridge have created a memory for computers that works like a brain
The new technology can fit 10 to 100 times more information on one device and process it in one place.
A team of scientists from the University of Cambridge has developed a new type of computer memory that could dramatically improve performance and reduce the power consumption of communications technology, which could consume nearly a third of the world’s electricity by 2030. The new memory processes data in a similar way to the synapses in the human brain and can store and process information in one place. The results of the study are published in the journal Science Advances.
The demand for energy in the world is growing every year due to the development of artificial intelligence, the Internet, algorithms and other data-based technologies. One reason for this is the shortcomings of current computer memory technologies, which require data to be moved between memory and processing, which takes both energy and time.
One possible solution is resistance-switched memory, which is capable of a continuous range of states, as opposed to traditional memory, which only has two states: one or zero. Such a memory can hold 10 to 100 times more information on a single device.
The scientists created a prototype device based on hafnium oxide, a material already used in the semiconductor industry. They found that adding barium to thin films of hafnium oxide leads to the formation of unusual structures that allow electrons to pass through and change the electrical resistance of the material.
The structures are like synapses in the brain: they can store and process information in one place, which makes them very promising for the rapidly developing fields of artificial intelligence and machine learning.
The technology was patented by the Cambridge Enterprise, a commercial arm of the university. Scientists are now working with industry to conduct more research on the possibilities of using materials. Because hafnium oxide is a material already used in the semiconductor industry, the scientists say it won’t be difficult to integrate it into existing manufacturing processes.