Intelligent Machines

Computing with Lasers Could Power Up Genomics and AI

The pace at which conventional chips improve is slowing, and these startups say optical computers are the answer.

As the cost of sequencing a person’s genome plummets, demand for the computing power needed to make sense of the genetic information is growing. Nicholas New hopes some of it will be worked on by a processor that parses data using laser light, built by his U.K. startup, Optalysys.

New says his company’s exotic approach to crunching data can dramatically upgrade a conventional computer by taking over some of the most demanding work in applications such as genomics and simulating weather. “The grunt work can be done by the optics,” he says.

Researchers have worked on the idea of using optics rather than electronics to process data for decades, with little commercial traction. But New argues that it is now needed because manufacturers such as Intel admit that they can’t keep improving conventional chips at the pace they used to (see “Moore’s Law Is Dead. Now What?”).

Manipulating laser beams encoded with data offers a shortcut on certain tricky computing operations.

Optalysys’s technology performs a mathematical function called the Fourier transform by encoding data, say a genome sequence, into a laser beam. The data can be manipulated by making light waves in the beam interfere with one another, performing the calculation by exploiting the physics of light, and generating a pattern that encodes the result. The pattern is read by a camera sensor and fed back into a conventional computer’s electronic circuits. The optical approach is faster because it achieves in a single step what would take many operations of an electronic computer, says New.

The technology was enabled by the consumer electronics industry driving down the cost of components called spatial light modulators, which are used to control light inside projectors. The company plans to release its first product next year, aimed at high-performance computers used for processing genomic data. It will take the form of a PCI express card, a standard component used to upgrade PCs or servers usually used for graphics processors. Optalysys is also working on a Pentagon research project investigating technologies that might shrink supercomputers to desktop size, and a European project on improving weather simulations.

French startup LightOn, founded earlier this year, has also built a system that uses optical tricks to process data being carried by laser light. The company is targeting a trick used in machine learning that compresses information by multiplying it with random data.

That technique is useful, but onerous for conventional computers working with large data sets, says Laurent Daudet, LightOn’s chief technology officer and a professor of physics at the Université Paris Diderot. His system gets the same effect more easily by exploiting the random scattering that naturally occurs when light passes through translucent material.

“Computers really suck at this, [but] you can use nature to do it instead,” says Daudet. He and LightOn’s two other founders published results last year showing how that works. The company  is working to show its system can perform other operations, too, and aims to put a new prototype online as a cloud service for others to experiment with, says Daudet.

The timing is good for these startups. Recognition that existing chip designs will no longer improve at the exponential pace they used to is making the industry more open to new ideas, says Horst Simon, deputy director of Lawrence Berkeley National Laboratory. “As conventional chips don’t get faster any more, alternative routes look more attractive,” says Simon.

Those alternative routes include quantum computers and chip designs specialized for AI, as well as optical computing systems. Simon doubts that optical technology will take over all kinds of computing, but says it might finally find a practical place. “There may be niche applications where its performance makes sense,” he says.

Tech Obsessive?
Become an Insider to get the story behind the story — and before anyone else.

Subscribe today
Manipulating laser beams encoded with data offers a shortcut on certain tricky computing operations.

Uh oh–you've read all of your free articles for this month.

Insider Premium
$179.95/yr US PRICE

More from Intelligent Machines

Artificial intelligence and robots are transforming how we work and live.

Want more award-winning journalism? Subscribe to Insider Plus.
  • Insider Plus {! insider.prices.plus !}*

    {! insider.display.menuOptionsLabel !}

    Everything included in Insider Basic, plus ad-free web experience, select discounts to partner offerings and MIT Technology Review events

    See details+

    What's Included

    Bimonthly magazine delivery and unlimited 24/7 access to MIT Technology Review’s website

    The Download: our daily newsletter of what's important in technology and innovation

    Access to the magazine PDF archive—thousands of articles going back to 1899 at your fingertips

    Special discounts to select partner offerings

    Discount to MIT Technology Review events

    Ad-free web experience

/
You've read all of your free articles this month. This is your last free article this month. You've read of free articles this month. or  for unlimited online access.