New AI chip mimics the human brain's capacity for split-second motor control — it solved problems using 10,000 times fewer calculations
A new brain-inspired device is designed to kick AI systems into gear only when they detect something unusual. Could it reduce dependence on power-guzzling data centers?
Scientists have created a new artificial intelligence (AI) chip that imitates the human brain's ability to quickly react to unexpected events, specifically the cerebellum. This part of the brain is responsible for coordinating balance and fine muscle control. The chip concentrates on unanticipated data, saving energy by ignoring routine information.
Researchers tested the device using electrocardiogram (ECG) data and found it could identify irregular heartbeats within one-fifth of a heartbeat with 98% accuracy, twice as fast as typical AI systems. Published in Nature Communications, the study suggests that this approach could lead to highly responsive, low-power AI systems for various applications, such as health monitors and autonomous robots.
The chip's innovative design integrates memory and processing components, known as a memtransistor, allowing for faster data transfer.
Written by urgent.news from Live Science's reporting — not their text. Machine-written — it may contain errors, so check the original before relying on it.
