For years researchers have been working on simulating brain synapses to create faster and more efficient computers.
Recently, a group of researchers from the Massachusetts Institute of Technology (MIT) has published a paper that describes a new type of chip that simulates brain synapses and offers improvement in performance and efficiency in comparison to all existing types.
The research team has come up with something called ‘memristors’ that have been created using silicon and a silver/copper alloy. These memristors can simulate brain synapses such that they can remember and recall data with high detail accuracy. Unlike typical transistors that can provide answers as true and false only, memristors provide a gradient of values functioning similar to a human brain.
According to the paper, the researchers ultimately aim at recreating a complex artificial neural network that can be run using memristors. Currently, these neural networks require significant GPU computing power to run. Using memristors, the team believes that they can localize neural networks in small devices, including Smartphones and cameras.
Moreover, using the concept of metallurgy, the team created super small chips that consist of tens of thousands of memristors. If this works out, it can completely change the way current supercomputers operate.
Although it is still a long way off, the team behind the project is hopeful that the breakthrough will lead to portable, artificial brain computers that can perform very complex tasks on the scale of today’s supercomputers.