Home → News → Brain-on-a-Chip Would Need Little Training → Full Text

Brain-on-a-Chip Would Need Little Training

By KAUST Discovery (Saudi Arabia)

April 21, 2021

[article image]

Researchers at King Abdullah University of Science and Technology (KAUST) in Saudi Arabia used a spiking neural network (SNN) on a microchip as a foundation for developing more efficient hardware-based artificial intelligence systems.

KAUST's Wenzhe Guo said SNNs mimic the biological nervous system and can process information faster and more efficiently than artificial neural networks.

The researchers created a brain-on-a-chip using a standard FPGA microchip and a spike-timing-dependent plasticity model, which allowed the neuromorphic computing system to learn real-world data patterns without training.

Compared to other neural network platforms, the brain-on-a-chip was more than 20 times faster and 200 times more energy efficient.

Guo said, "Our ultimate goal is to build a compact, fast and low-energy brain-like hardware computing system."

From KAUST Discovery (Saudi Arabia)

View Full Article

Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found