PAPER ALERT: Researchers from @utexasece are re-imagining computers to think like the human brain and they’ve made a new device discovery. They used magnetic materials to engineer nano-devices that behave like synapses in the brain, which connect neurons to each other. The synapses they created can be adjusted by manipulating their geometry, which allows the synapses to be tuned for a variety of potential computing applications. Brain-like, or neuromorphic devices represent a new computing paradigm that can perform complex tasks that the human brain is good at, like image processing, and they learn as they go. Magnetic materials fit many of the major requirements for brain-like computing because of their ability to switch not just between 0s and 1s but to other levels, and stay that way when they are off. The researchers recently published a similar paper on neuromorphic synapses using biocompatible materials, which are important for medical uses, but not fast enough for next-gen computers. This paper is focused more on devices with the speed and energy to compete with state-of-the-art computing technologies, as we think of how to design computers for AI.
Click here to read!
