Knowm, a startup launched only six months ago, has announced this week that two new variations on the Knowm memristor are available for purchase. Knowm is also selling raw device data that can be used by researchers to develop their own memristor models.
Memristors are an electrical component that, much like the synapses in the brain, can remember previous connections. Neurons learn by sending pulses through synapses, which strengthens the neural pathway, or not sending pulses which causes the synapses to atrophy. In a memristor, the neural pathway is emulated through resistance – an electrical connection that is used frequently will have a lower resistance; one that is rarely used will have a higher resistance. Knowm’s memristors can learn and retain information based on characteristics of data flow, data frequency, and current.
Read the entire article directly on The Stack website at: https://thestack.com/world/2016/01/21/knowms-new-memristor-improves-efficient-ai-processors/
Knowm’s new memristor improves efficient AI processors