Prototype optical AI accelerator chip debuts

Prototype optical AI accelerator chip debuts

Technology News |
By Rich Pell

CEO Yichen Shen is quoted saying: “We are very pleased to reveal our working optical chip AI computing system. Our prototype … is 100,000 times faster than the system demonstrated in our Nature Photonics paper and a fraction of the size.”

Other claims made for the chip is that latency is improved up to 10,000 times compared with traditional hardware and power consumption is orders of magnitude lower.

Lightelligence was founded in 2017 and has received seed funding of about $10.7 million in a funding round led by Baidu Ventures and a group of US semiconductor executives. Baidu is the leading Chinese internet services company.

In the Nature Photonics paper Shen reported the use of a programmable nanophotonic processor featuring a cascade of 56 programmable Mach–Zehnder interferometers to implement neural networks for vowel recognition. The prototype has been used to implement the MNIST machine learning model to recognize hand-written digits, the report said.

Although optical processing has been studied academically for many decades it has yet to find commercial application. In theory, optical processing supports high parallelism at much lower power than electronic computation.

Now MIT is supporting two spin-offs to pursue optical processing. Lightelligence is in competition with Lightmatter Inc. (Boston, Mass.) which recently raised $22 million in a round led by Google Ventures and which brought its funding to $33 million.

Nature Photonics article

News articles:

Google invests in photonic processor startup

MIT spin-off raises funds for optical processor

Nanoscale photonics switch could enable 10-100 Terabit/s data transfer

Solid-state lidar maker opens factory

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles