Behind the scenes at Labs
cancel
Showing results for 
Search instead for 
Did you mean: 

HPE DISCOVER: Do we have to wait for quantum computers?

Curt_Hopkins

rayfar (Custom).jpgRay Beausoleil

Do we have to wait for quantum computers?

“Depending on the problem we’re trying to solve, no,” says HPE senior fellow Ray Beausoleil, who heads up Hewlett Packard Labs’ Large-Scale Integrated Photonics research group.

Quantum computing has succeeded in capturing the public imagination, but its effect on enterprise users would be minimal, at best. Typically, HPE customers want to start with massive amount of data and analyse it, in order to learn from it. There is an integral problem with quantum computers making this standard computation.

Beausoleil will discuss this and other roadblocks as well as innovative classical alternatives to quantum computing in the HPE Discover session, “Do we have to wait for quantum computers?” (IS5090)

But we’ll give you a few insights to arm yourself for the session right here and now.

Classical computers use deep learning and other methods to sift through these large databases. In order for a quantum computer to do this same action, the database would need to be copied into a quantum register, which is a difficult thing to do. But the problems don’t stop there.

“Nature is doubly unfair because you can’t convert a classical database into quantum memory and then store the quantum database for later use,” says Beausoleil. This “no cloning theorem” says you can’t copy a quantum state. The implication of that? There can be no such thing as a quantum hard drive. “Deep learning is likely to remain a job for classical computers.”

Other reasons that quantum computers are unlikely to provide a valuable assist to enterprise customers include the fact that there is no quantum computer large enough yet to solve an interesting problem. The field as a whole is trying hard to demonstrate “quantum supremacy,” where a quantum computer outperforms a classical computer on one of a carefully chosen set of computational problems. Unfortunately, the answers to these problems are not particularly interesting.

“People are working their asses off to get a quantum computer to do anything at all! Quantum Supremacy would be an important engineering milestone that shows that the community is headed in the right direction” says Beausoleil.

The most likely first use of quantum computers will be to simulate other quantum systems. In other words, they may be used to design new materials, develop drugs, and answer longstanding questions in theoretical physics; none of which are likely to be highly valued in the C-suite.

One possible area of interest that quantum computers may prove useful in exploring are NP-hard problems, such as the traveling salesmen problem. But the wait for quantum computing is likely to be too long for people dead set on solving problems in the here and now. For instance, chaos computing has proven helpful in NP-hard problems and on a much shorter time frame, and optical computing may do the same within the next few years.

“For the foreseeable future,” says Beausoleil, “the wise enterprise user will be better served focusing on solutions like HPE’s Dot Product Engine, an accelerator designed to help deep learning work go much faster and a lower cost.”

Beausoleil will outline the promise and problems of quantum computing, and the benefits of innovation to classical computing at the session below.

Do we have to wait for quantum computers?

Session ID: IS5090

Date: Thursday, June 21

Time: 11:30 – 12:00 pm

Session Owner: Rebecca Lewington

Speaker: Ray Beausoleil

Register for HPE Discover 2018 in Las Vegas, June 19-21

Photo by Rebecca Lewington

0 Kudos
About the Author

Curt_Hopkins

Managing Editor, Hewlett Packard Labs