Try as we might to build and upgrade better computers and components, the most advanced and capable computer in existence is the human brain. With billions of synapses with various connections sending signals to each other, the brain is a highly efficient and parallel machine no computer can match. Naturally though, when told something is better than what you have, you try to match it, and that is what researchers at Harvard University are working on as they develop a synaptic transistor.
Though it may be hard to imagine that every memory and thought you have had exists just as a network of neurons, that does appear to be the case. New experiences can cause new connections to form between neurons, and as the experiences are repeated, those connections are strengthened so signals can travel faster along them. In the typical transistor, the connections are static, so to emulate the brain, the Harvard researchers have developed a transistor with adaptable connections. The device is comprised of a nickelate semiconductor between two platinum electrodes, next to some ionic liquid. By applying an electric field to the liquid, ions are driven into the nickelate, changing its conductance and altering the connection between the electrodes.
The current device is just a proof of concept for the synaptic transistor, but the researchers already have an idea of how to improve its performance. By moving the ionic liquid closer to the device, the transistor should be able to respond faster, so the researchers are looking to work with some microfluidics experts to achieve this greater speed.
Source: Harvard University