Sydney researchers build ultra-compact AI chip operating at speed of light

Researchers at the University of Sydney have built an ultra-compact Artificial Intelligence (AI) chip that is able to make calculations using the power of light, at the speed of light. Their study, published in the journal Nature Communications, demonstrates how AI models can be designed into nanoscale photonic structures that manipulate light to perform the mathematical operations required for machine learning.

The nano photonic chip prototype is built completely in-house at the university’s Sydney Nano Hub.  To validate the technology, the researchers trained the nanophotonic chip to classify more than 10,000 biomedical images such as breast, chest and abdomen MRI scans. In simulations and experiments, the nanophotonic neural network achieved approximately 90 to 99 percent classification accuracy. 

Traditional computer chips use electricity to process information, which means moving electrons through wires. This produces heat. Instead, the nano photonic chip prototype uses light which travels through materials without electrical resistance, so does not generate heat in the same way electricity does. As the light passes through the nanostructures within the chip prototype, these structures themselves perform the calculation automatically. 

The researchers say the prototype could play an important role in developing more energy-efficient AI hardware as global demand continues to grow, potentially lowering the overall energy footprint of future computing systems. The technology uses light to operate instead of electricity, compared to current data centres that rely on massive amounts of water and energy to power them.

 “We’ve re-imagined how photonics can be used to design new energy efficient and ultrafast computer processing chips,” said Professor Xiaoke Yi from the University’s School of Electrical and Computer Engineering and director of the Photonics Research Group. 

“Artificial intelligence is increasingly constrained by energy consumption. This research performs neural computation using light, enabling faster, more energy-efficient and ultra-compact AI accelerators.”

The researchers say the advantage of using photonics is the computation is much faster, taking place at the speed of light. The prototype performs calculations on the picosecond timescale, or trillionths of a second – the time it takes light to pass through the nanostructure. 

The technology offers a pathway to sustainable AI infrastructure capable of supporting the growing demands of computing without proportional increases in power consumption, says the university in a media release. 

For more than a decade, the Photonics Research Group at the University of Sydney has been studying how to push the limits of photonics. This includes using photonics to tackle challenges in wireless communications and advanced sensing technology which can detect and measure chemical or biological traces in the environment. 

Following the successful testing of the nanophotonic chip prototype, Professor’s Yi’s team is now working on advancing the technology toward larger-scale photonic neural networks.