Email Print Share
December 1, 2022

Learning with Light

Researchers at MIT have developed a new edge computing architecture that could revolutionize smart devices.

Credit: U.S. National Science Foundation


Learning With Light

Anytime you ask a smart device a question or search for directions, it takes several seconds for a response. That latency is due to the memory needed to answer your question being housed at a data center. Your query may travel hundreds of miles away to be processed before it returns an answer to you.

But what if we could have a response in real-time? We'll explore in the U.S. National Science Foundation's "Discovery Files."

Although our smart devices are increasingly powerful, they are limited in the amount of processing that they can perform locally by their size, weight, and power. Most of the heavy lifting is done in the cloud, the factory sized servers that use massive amounts of power in data centers around the globe.

At MIT, NSF-supported researchers have created a new system that computes directly on these devices at teraflop rates.

By storing the weight or computational parameters on a centralized server that periodically connects to a smart transceiver via light waves, they have created a scalable, edge computing architecture called Netcast that showed a hundredfold improvement in energy efficiency when comparted to other models.

The Latency free processing method could allow self-driving cars to make decisions with just a fraction of the energy currently required, process live video over cellular networks, or improve security as transmission of personal data to a central location would no longer be required.

To hear more science and engineering news, including the researchers making it, subscribe to "NSF's Discovery Files" podcast.


Images and other media in the National Science Foundation Multimedia Gallery are available for use in print and electronic material by NSF employees, members of the media, university staff, teachers and the general public. All media in the gallery are intended for personal, educational and nonprofit/non-commercial use only.

Images credited to the National Science Foundation, a federal agency, are in the public domain. The images were created by employees of the United States Government as part of their official duties or prepared by contractors as "works for hire" for NSF. You may freely use NSF-credited images and, at your discretion, credit NSF with a "Courtesy: National Science Foundation" notation.

Additional information about general usage can be found in Conditions.