Neuromorphic computing, which uses chips that mimic the behavior of the human brain using virtual “neurons,” is growing in popularity thanks to high-profile efforts from Intel and others. Now, a team of researchers led by the Swinburne University of Technology in Melbourne is announcing that it has demonstrated what it’s calling “the world’s fastest and most powerful optical neuromorphic processor for artificial intelligence.”

The processor is capable of achieving 10 trillion operations per second, which the researchers claim is 1,000 times faster than “any previous processor.” Specifically, the researchers report, the chip excels at processing ultra-large scale images with remarkable speed, achieving full facial image recognition.

“This breakthrough was achieved with optical micro-combs,” said David Moss, director of the Optical Sciences Centre at Swinburne and co-lead of the research. The same technology, Moss said, was used to achieve the all-time record-setting internet speeds reported by Swinburne and partnered institutions last May. 

“In the 10 years since I co-invented them, integrated micro-comb chips have become enormously important and it is truly exciting to see them enabling these huge advances in information communication and processing,” Moss said. “Micro-combs offer enormous promise for us to meet the world’s insatiable need for information.”

Essentially, the small, light and – crucially – cheap micro-combs each contain hundreds of high-quality lasers, enabling much faster transfers of information: in the fiber internet use case, for example, a single microcomb supplanted 80 individual lasers. For the neuromorphic chip use case, a single micro-comb-enabled processor was used to simultaneously assess the data across multiple dimensions, a task the researchers say would have required tens of thousands of parallel processors when performed using the comparably powerful Google TPU.

“This processor can serve as a universal ultrahigh bandwidth front end for any neuromorphic hardware – optical or electronic based – bringing massive-data machine learning for real-time ultrahigh bandwidth data within reach,” said Xingyuan (Mike) Xu, a research fellow at Monash University who co-led the research. “We’re currently getting a sneak-peak of how the processors of the future will look. It’s really showing us how dramatically we can scale the power of our processors through the innovative use of microcombs.”

The research was published in Nature under the title “11 TOPS photonic convolutional accelerator for optical neural networks.” The article was written by Xingyuan Xu, Mengxi Tan, Bill Corcoran, Jiayang Wu, Andreas Boes, Thach G. Nguyen, Sai T. Chu, Brent E. Little, Damien G. Hicks, Roberto Morandotti, Arnan Mitchell and David J. Moss.