Unlocking the Hardware Gap: Exploring the Potential of Optomemristors for Artificial Neural Networks

Using Optomemristors To Light Up Artificial Neural Networks

Research in artificial intelligence and machine-learning hardware has focused on building photonic neurons and synapses, and combining these to perform fundamental neural-type processes. Complex processing methods in the human brain, such as dendritic computing and reinforcement learning, are more difficult to reproduce directly in hardware.

The \”hardware gap\”, as it is called, can be closed by developing an \”Optomemristor\”, a device that can respond to multiple electronic and photonic signals at once. Diverse biophysical mechanisms govern the functions and synapses of the mammalian neuronal brain.

It is possible to simulate the multi-factor computations of mammalian minds using chalcogenide thin films. This technology uses both electrical and light impulses, and consumes very little energy.

Source:

Using Optomemristors To Light Up Artificial Neural Networks