I'm admittedly a novice in this field, but I have read that, while the D-wave (one) is an interesting device, there is some skepticism regarding it being 1) useful and 2) actually a 'quantum computer'.

For example, Scott Aaronson has expressed multiple times that he is skeptical about whether the 'quantum' parts in the D-wave are actually useful:

It remains true, as I’ve reiterated here for years, that we have no direct evidence that quantum coherence is playing a role in the observed speedup, or indeed that entanglement between qubits is ever present in the system.

Exerpt from this blog.

Additionally, the relevant Wikipedia section on skepticism against the D-wave is a mess.

So, I ask:

  1. I know that D-wave claims to use some sort of quantum annealing. Is there (dis)proof of the D-wave actually using quantum annealing (with effect) in its computations?

  2. Has it been conclusively shown that the D-wave is (in)effective? If not, is there a clear overview of the work to attempt this?

  • 1
    that blog post is quite old, you should have a look at the more recent ones, e.g. scottaaronson.com/blog/?p=3192 and scottaaronson.com/blog/?p=2555 – glS Mar 16 at 17:53
  • @glS Thanks. I just looked up the first blogpost related to D-wave that I could find. I mean, the man talks about D-wave to the point he gets sick of it. Still, I think the quote is relevant. – Discrete lizard Mar 16 at 18:00
  • 1
    I'd be interested in seeing more answers than just one from someone who seems to have worked with D-wave in the past and whose advisor showed that D-wave does have quantum effects; i.e., while I don't believe Andrew O. is trying to be deceptive, it would be interesting to see another perspective from someone who can "view from a distance", so to speak. – heather Mar 17 at 16:23
  • 2
    Discretelizard, @heather I added some references to papers that suggest the D-Wave One is a simple (non-quantum) thermal annealer, and another paper that suggests the finite temperature as the reason. – Andrew O Mar 18 at 20:37
  • @Discretelizard: If I may clarify your question, do you really care whether the D-Wave One (a now ~7-year-old system) is proven quantum, or whether any D-Wave system (including later models) has been proven quantum? – spreinhardt Mar 21 at 1:17
up vote 23 down vote accepted

There is still a search for problems where the D-Wave shows improvement over classical algorithms. One might recall media splashes where the D-Wave solved some instances $10^8$ times faster than a classical algorithms but forgot to mention that the problem can be solved in polynomial time using minimum weight perfect matching.

Denchev showing $10^8$ speedup https://arxiv.org/abs/1512.02206

Mandra using MWPM https://arxiv.org/pdf/1703.00622.pdf

There is some evidence that there are indeed some quantum effects used by the D-Wave. Notably a study by Katzgraber et al. that compares the D-Wave with simulated annealing and the effects of reducing barrier thickness in the energy landscape (to make tunneling more probable). In Fig. 5 of the following paper the barrier thickness is reduced and the D-Wave shows improvement on the class of problems while Simulated Annealing shows no improvement.

https://arxiv.org/pdf/1505.01545.pdf

Full disclosure: Katzgraber was my PhD advisor so I am most familiar with his work.

On the other hand, there have been a few papers on the topic of the D-Wave being a simple thermal annealer with no quantum effects, notably the papers by Smolin although they are a bit dated now.

https://arxiv.org/pdf/1305.4904.pdf

https://arxiv.org/pdf/1401.7087.pdf

More recently Albash et al. discussed the finite temperature as a reason for quantum annealers not functioning competitively.

https://arxiv.org/pdf/1703.03871.pdf

  • 2
    While a $10^8$ speedup is nice, a constant speedup on a problem with a known polynomial algorithm seems a bit meager compared to the theoretical 'promises' of quantum computation. I like your answer, but I do hope someone posts another answer showing work on the side of showing 'ineffectiveness' and 'non-quantumness' (although I guess the latter is hard to show), so that we have multiple views until research gives us more conclusive results. – Discrete lizard Mar 16 at 19:53
  • Didn't you work with DWave, on a benchmark paper? – heather Mar 17 at 1:20
  • @heather Yes. I have collaborated with a colleague at D-Wave, Firas Hamze, on several projects; designing problems with variable complexity, best-case performance on quantum-annealers due to noise, and the benchmarking paper listed above. – Andrew O Mar 17 at 8:53
  • 2
    @Discretelizard I would hardly characterize a hundred-million speedup "meager". Complexity theory is way too obsessed with infinite size scaling limits. What matters in real life is how fast you get the answer to your problem. – DanielSank Mar 18 at 21:25
  • 1
    @tparker I never put forth any opinion about whether or not DWave could be a Pentium in an apples-to-apples comparison. I simply commented that a hundred million speedup would be worthy of some note. – DanielSank Mar 28 at 15:53
  • Is there proof that the D-wave (one) is a quantum computer and is effective?

D-Wave Video - Offers an explanation of: "How do we know ...": https://youtu.be/kq9VqR0ZGNc

One analogy you might make with the D-Wave One, an adiabatic ('analog') computer, is to the "south-pointing chariot" or the "Antikythera mechanism".

A lengthy explanation is offered in this Ars Technica (Wired) article: "Going digital may make analog quantum computer scaleable":

  • "... They pretty much all fall into two categories. In most labs, researchers work on what could be called a digital quantum computer, which has the quantum equivalent of logic gates, and qubits are based on well-defined and well-understood quantum states. The other camp works on analog devices called adiabatic quantum computers. In these devices, qubits do not perform discrete operations, but continuously evolve from some easily understood initial state to a final state that provides the answer to some problem" (end quote), or quantum annealing.

  • "Adiabatic quantum computers are inherently analog devices: each qubit is driven by how strongly it is coupled to every other qubit. Computation is performed by continuously adjusting these couplings between some starting and final value. Tiny errors in the coupling—due to environmental effects, for instance—tend to build up and throw off the final value.".

  • "Digital quantum computing, which uses logic operations and quantum gates, offers the possibility of error correction. By encoding information in multiple qubits, you can detect and correct errors. Unfortunately, digital qubits are delicate things compared to those used in adiabatic quantum computers, and the ability to ...". (Go read the article if you don't want a condensed version).

  • "What about a hybrid approach? That's the question asked by a international group of researchers in a recently-published paper in Nature. They’ve tested a system where the computation is performed by qubits that were operating as an adiabatic quantum computer, but with connections between the adiabatic qubits is controlled via a digital network of qubits. This allows the benefits of scale and flexibility that you get from adiabatic quantum computing, while also making the connections less susceptible to noise.".

So, yes. It is a computer and uses quantum methods.

Adiabatic quantum computation (AQC) is a form of quantum computing which relies on the adiabatic theorem to do calculations1 and is closely related to, and may be regarded as a subclass of, quantum annealing.

Another analogy, probably as unfair as the last, is that AQC is a one-trick-ponyism. It's limited in what it can do, but it does it quickly and well.

  • So, I ask:

    I know that D-wave claims to use some sort of quantum annealing. Is there (dis)proof of the D-wave actually using quantum annealing (with effect) in its computations?

    Has it been conclusively shown that the D-wave is (in)effective? If not, is there a clear overview of the work to attempt this?

There is proof that it is effective when used correctly for doing what it was designed to do:

"Blockchain platform with proof-of-work based on analog Hamiltonian optimisers" by Kirill P. Kalinin, Natalia G. Berloff, 27 Feb 2018.

University of Cambridge, "Polariton Graph Simulator (Optimizer): an analog Hamiltonian simulaton", Natalia Berloff.

"Performance of quantum annealing hardware" by Damian S. Steiger; Bettina Heim, 22 October 2015.

There exists important backers and some skeptics of D-Wave.


Address concerns expressed in comments - Update: 19 March 2018:

Here is an article from Nature.com entitled: "Triode for Magnetic Flux Quanta" which explain the use of Abrikosov vortices to hold quantized information bits, further clarified (or not) in the article: "Single Abrikosov vortices as quantized information bits".

An oversimplified analogy is that the quantum qubits are (not at all) like magnetic core memory, the difference is:

  • A single magnetic core holds a binary digit, a bit, (like a fraction of a letter in a book, so you would use 8 bits to represent more than just a letter but all of the ASCII Alphabet, letters digits and control codes). A bit would have to be in one state or the other.

  • A qubit, by utilizing quantum mechanics, allows the qubit to be in a superposition of both states at the same time, a property that is fundamental to quantum computing. A qbit can be in one state, the other, or both; think of it as trinary on steroids, because qubits can perform two calculations simultaneously (and that's why they are both comparable and incomparable, a superposition of both states; a new way if thinking).

Look at this image of a magnetic memory and a quantum processor - quite different from an x86 processor:

Which is which?

A simple explanation of the relevance and degree of proof is offered in this video by D-Wave called: "D-Wave Lab Tour Part 3 (of 3) - The D-Wave Processor".

https://www.youtube.com/watch?v=AGByZoYUlU0

  • 1
    "So, yes. It is a computer and uses quantum methods.". Ah, but I specified (albeit parenthetically), for proof that 'quantumness' is used with effect by the computer. If we ignore effective usage, I could just tape a box with some qubits in it to my laptop and call it a quantum computer! Also, I'm afraid I'd prefer having academic sources for this proof, not news sites (the link is appreciated though, it merely doesn't meet my standards for proof) – Discrete lizard Mar 18 at 9:32
  • 1
    The final references look nice, but it would be even better if you briefly describe why they're relevant here. Thanks! – Discrete lizard Mar 18 at 9:34
  • The 'news site' was presented to provide a layperson 's explanation of degrees or types of quantumness as to your second comment it dismisses ~40% of your first comment, the very sources you request. Did you read a bit, fire off a comment, read some more, fire off another comment ... -- don't expect replies if you do that. – Rob Mar 18 at 9:53
  • 1
    Oh wait, I think I understand you. While my comments were made a couple of minutes apart, both were made after reading you answer completely, although I haven't yet read all the material from your links. – Discrete lizard Mar 18 at 18:04
  • Let me make a brief remark. I know you have rejected my edit that removed the 'timestamp'. The reason I removed it is because it is not needed: the exact time of all edits is already tracked in the public history of you question. Additionally, we like answers to be 'a single whole', without notes such as 'edit' and other parts that are not needed. So, I ask you to reconsider and remove the 'timestamp' yourselves. Thanks. – Discrete lizard Mar 22 at 8:24

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.