Brain function that occurs in metallic nanowire networks, The research team led by NIMS was able to develop a neuromorphic network consisting of several metal nanowires.
Using this network, the team can produce electrical properties that are similar to high-level brain functions that are unique to humans, e.g.
The development of artificial intelligence (AI) techniques has made rapid progress in recent years and began to affect our lives in various ways. Although AI processes information similar to the human brain, the mechanism used by the human brain is still unknown.
Basic brain components such as neurons and junctions (synapses) have been examined in detail.
However, many questions about the whole brain need to be answered. For example, we do not fully understand how the brain performs functions such as memorization, learning and forgetting, and how the brain becomes alert and returns to peace.
The research team recently built a complex network for the brain by integrating several silver (Ag) wires that were coated with a polymer insulation layer (PVP) of about 1 nanometer.
The transition between two nanowires forms a variable resistance element (eg, synaptic element) that behaves like a synaptic neuron. These nanowire networks, which contain a large number of synaptic elements that interact with each other, form neuromorphic networks.
When voltage is applied to a neuromorphic network, it appears that it is “difficult” to find the optimal current path (ie the most efficient path with electricity).
The research team measured the process of forming, storing, and deactivating the current paths when electricity flows through the network, and found that, like the process of storing, learning, and forgetting the human brain, this process always oscillates when they flow.
The research team is currently developing devices such as the brain using neuromorphic tissue material. The team intends to design the storage device to function. Basically, a different principle is used compared to computers today.
The team also hopes that this research will facilitate understanding of information processing mechanisms in the brain.