Neuromorphic Chips and Spiking Neural Networks Aim to Cut AI's Energy Use

Conventional AI relies on a computing design dating to 1945, in which processors constantly shuttle data to and from separate memory, consuming large amounts of power. Researchers are developing neuromorphic computing, which takes inspiration from the brain, to address this inefficiency. These chips use spiking neural networks that process information through timed electrical spikes and operate asynchronously, activating only when changes occur. The approach co-locates memory and computation in memristor-based synapses, reducing energy demand and latency. Potential applications include onboard automotive and aerospace systems and low-power brain-computer interfaces.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)
Log in to join the discussion and vote.
Log in