Friday, 20 October 2017

U of A Lecture on Quantum Computing (D-Wave Systems)



U of A Lecture on Quantum Computing (D-Wave Systems)

The Speaker

We went to a lecture the other week (September 28, 2017) about developments in quantum computing, put on by the University of Alberta, for the Physics and Research Symposium and Public Outreach program.  The lecture was given by Dr. Emile Hoskinson, an experimental physicist at D-Wave Systems (located in Burnaby, British Columbia, near Vancouver), and thus focussed on that corporation’s “spin” (no quantum mechanics pun intended) on quantum computing.  Dr. Hoskinson did his undergrad at UBC, and his PhD at Berkeley. He went to high school just down way from the U of A, though, at Archbishop MacDonald High School, so he had a local connection.

He described his job as “to design, process, test, calibrate, and run experiments to evaluate performance of the D-Wave supercomputers”.  He also described his workplace as “one of the coolest places there is”, a riff on the fact that quantum computing is just plain cool, in the vernacular sense of the term, and that the process itself operates at near absolute zero, for reasons described below.

I should note that the talk was pitched at a general audience, so he intended it to be understandable, yet not dumbed down.  I think he succeeded in that objective, and I sensed that the audience would agree with that.  He also did a physics colloquium during his visit – presumably that was a more technical presentation.

I should also note that the public talk didn’t go into quantum theory in any depth – quantum tunnelling and superpositions were the main aspects of quantum theory that were touched upon.  So, it no doubt helped to have had some acquaintance with quantum theory, to get a better handle on the talk.  I have some background – basically lots of reading, and what mathematical/technical understanding that an undergrad in physics will confer.  But, obviously, to understand the technology at a deeper level would require a significant immersion in the subject.  The D-Wave site has plenty of description and documentation that the interested reader can peruse.

Quantum Computing Progress


There are several approaches to using quantum phenomena for computing, and D-Wave specializes in one particular approach, but more about that a bit further on.  It should be noted that the D-Wave approach has both academic and commercial aspects.  On the commercial side, buyers have included such outfits as NASA, Google and Lockheed Martin, and some 150 patents have been filed.  On the academic side, there have been some 90 peer reviewed papers written, relating to the technology. 

D-Wave One, their first commercial quantum computer was released in 2010; it had 128 Qbits of quantum processing capacity.  D-Wave 200Q is the most recent release, in 2017; it has 2000 Qbits of capacity.  The capacity of these computers has followed “Moore’s Law” like trajectory, with the number of Qbits increasing from 4 in 2004 (early research) to about 10,000 in 2018 (20,000 is possible in the next release).

Here’s my graph of that, from some things said during the talk (note that it is not official by any means, and I only have 4 data points).  I make the doubling time to be about 1.25 years.

I should note that a Qbit is something like a “bit” in regular computing.  However, where a regular bit can be in two states (and thus naturally leads to binary Boolean logic), a Qbit can exist in State 0 (off), State 1 (on) or a superposition of the two.  You can now meditate upon Schrodinger’s Cat, to consider the ramifications of such a device.  Plus, think a bit about quantum tunnelling.  As will be explained a bit later (to the extent it can be explained),  quantum tunnelling is probably the key phenomenon that D-Wave’s make use of.





The Quantum Computer

So, what is a quantum computer, as operationalized by D-Wave?  Visually, as he demonstrated in his presentation, it looks pretty much like a big black box. 

The Black Box


The black box has two main purposes:

  • It acts as a Faraday Cage, keeping stray electromagnetic signals away from the quantum chip, which does the quantum part of quantum computing.  Stray signals can interfere with the delicate process of maintaining quantum superpositions, which, of course, is the key to a quantum computer’s advantage over regular computing.

  • It contains the hardware necessary to produce the low temperatures at which the quantum chip operates.  Again, this has to do with maintaining a quantum superposition state for useful lengths of time – thermal agitation at the molecular level (i.e. heat) will also interfere with this.

  • The operating temperatures for the quantum chip are about 15 milli-Kelvins, or about 15 thousands of a Celsius degree above absolute zero.

  • The computer’s temperature is lowered via multiple stages, with each stage dropping the temperature more and more.   The final stage contains the quantum chip.


Fridge Wiring



The quantum chip looks pretty normal, somewhat like a GPU processer used in graphics applications.  It actually is based on small, but still macroscopic devices which create superconducting current loops.  Thus, the need for near absolute zero temperatures.  The current can flow in either of the two directions around the loop, creating a digital one or zero.  But it can also quantum tunnel between these states, which is the key to quantum computing, of course.  The direction and amplitude of the current in these loops is altered by applying a magnetic bias to the loop.  In this respect it sounded to me somewhat like “core” memory in the old mainframes of the past era, but with a superconducting quantum twist to it.

Quantum Chip



Note that the computer also has a conventional front end, as well as the quantum chip back end.  The quantum computer, as realized with this technology is only productive for certain types of problems, that it is optimized for.  These tend to be algorithms that don’t scale up to huge sizes well.

 

Quantum Computer Applications


An example given was essentially as sort of permutation problem, which has a huge search space as it is scaled up.  Finding the most efficient solution to a logistical problem or a consumer preference optimization might come to mind – problems in finding correlations in genetics were another example mentioned.

Suppose one was searching for an optimal solution to such a permutation problem.  Normally, finding the global minimum in such a search space would soon get out of hand, as the problem would grow exponentially as it is scaled up.

But, with clever design of the quantum chip, the chip can be made in such a way that it mimics the physical or conceptual problem.   The chip can then quantum tunnel to get out of a local minimum, which can be a huge problem in conventional computing, requiring computing time and resources that are not practical (it sounds like a gradient descent problem, a key aspect of many AI algorithms).  However, the quantum chip will evolve to a ground state solution, via quantum mechanics.  If the chip has been designed to mimic the physical problem, this can give the solution to the problem.



Note that this can involve a lot of custom design of the chip, to fit the specified problem.  Obviously, not all interesting and useful problems in computing can be solved via this technology.  More general purpose quantum computers are being explored, though they are still in the early stages.

In some ways, the D-Wave quantum computer reminded me of analogue computers, in the sense that the hardware is built to mimic a physical problem of interest.  In the past, if I recall correctly, this was a method for solving differential equations.  Basically, one designed a circuit that corresponded to a particular differential equation, and solved the equation via analysis of the corresponding circuit’s behaviour.

 

Richard Feynman on Quantum Computing


Dr. Hoskinson noted that Richard Feynman once said about the possibilities of quantum computing: 
And I'm not happy with all the analyses that go with just the classical theory, because nature isn't classical, dammit, and if you want to make a simulation of nature, you'd better make it quantum mechanical, and by golly it's a wonderful problem, because it doesn't look so easy.
International Journal of Theoretical Physics, VoL 21, Nos. 6/7, 1982 Simulating Physics with Computers Richard P. Feynman

This is pretty mind bending stuff, so I would also add that he once quoted as saying: 
 "If you think you understand quantum mechanics, you don't understand quantum mechanics."

I think many of us can agree with him on that point, and that goes double for understanding quantum computers.  Nonetheless, the lecture was very informative, entertaining and engaging.

Sources:



------------------------------------------------------------------------------------------------------------

Now that you have read of some cutting edge science, you should consider reading some Science Fiction.  How about a short story, set in the Arctic, with some alien and/or paranormal aspects.  Only 99 cents on Amazon.

The Magnetic Anomaly: A Science Fiction Story

“A geophysical crew went into the Canadian north. There were some regrettable accidents among a few ex-military who had become geophysical contractors after their service in the forces. A young man and young woman went temporarily mad from the stress of seeing that. They imagined things, terrible things. But both are known to have vivid imaginations; we have childhood records to verify that. It was all very sad. That’s the official story.”



No comments:

Post a Comment