back to article Useful quantum computers will be impossible without error correction. Good thing these folks are working on it

Boffins from America's standards-setting body NIST, the University of Maryland (UMD), and the California Institute of Technology believe there's a way to make quantum computers correct many of their own errors – which would help overcome one of the major design challenges for such devices. Physicists Simon Lieu, Ron Belyansky …

  1. Mike 137 Silver badge

    The fundamental problem?

    The fundamental problems are firstly that quantum phenomena are essentially statistical, so a quantum logic block might legitimately be called a "maybe gate", and secondarily that minute levels of interference ("noise") are likely to be sufficient to completely modify results. Under these circumstances error correction using additional bits suffers from diminishing returns, as the additional bits are also both statistical and subject to interference. So recursively, you might need additional bits to correct (or at least validate) the error correcting bits - ad infinitum. Consequently, as at some point we have to stop adding layers of correction, there will always be a practical limit to the consistency of results from a quantum computer.

    1. Paul Kinsler

      Re: The fundamental problem?

      Quantum phenomena are not "essentially statistical". Quantum systems often behave as if they suffer statistical variations, typically as a result of interaction with a much larger environment, but this behaviour is not fundamental to QM. QM can get along quite nicely all by itself with unitary evolution, although such a purist approach makes understanding things like measurement philosophically tricky.

      More pedantically (yes, more!), I wouldn't use "interference" when talking about noise in quantum systems - "interference" has a technical use, and although in everyday and/or colloquial contexts "interference" and "noise" might be of sufficiently similar meaning, such relaxed usage is a recipe for confusion in quantum physics.

      1. Claptrap314 Silver badge

        Re: The fundamental problem?

        I know very little of QM, but he sounds like he doesn't understand the basics of error correction.

        1. el kabong

          There are no basics of quantum error correction

          Conjectures is all we have now.

          Hopefully, someday, there will be.

          Or not.

          1. circusmole

            Re: There are no basics of quantum error correction

            ..or both.

      2. CrackedNoggin Bronze badge

        Re: The fundamental problem?

        In a perfectly isolated (up to final observation) QC, to get the final result back into the real world, the system needs to be observed, and only one of it's simultaneous states (among a distribution of states) will be observed. Which one is observed is a statistical sampling - is it not? If computing a very "hard" problem like breaking a cryptographic key, that distribution should be very close to a singularity to be of any use. However if it is a soft problem like "generating a random distribution" (a.k.a. "proving quantum supremacy"), then it needn't be nearly a singularity.

        A separate issue is how decoherence affects the QC. That will skew the distribution to a greater or lesser degree, including collapsing it to a singularity (the wrong one). If performing the the quantum many times, then there is another meta-distribution of skewed distributions (skewed by the decoherence noise). Which again is statistics.

        1. Paul Kinsler

          Re: Which again is statistics.

          I was quite careful - I was taking issue solely with the statement "quantum phenomena are essentially statistical"; and not making a statement about any particular implementation of some device and any assumed measurement process, whether in the context of quantum computing, or not.

      3. amanfromMars 1 Silver badge

        Re: The Fundamental Problems for Solution are within the Gift of AI Bypass Operations

        More pedantically (yes, more!), I wouldn't use "interference" when talking about noise in quantum systems - "interference" has a technical use, and although in everyday and/or colloquial contexts "interference" and "noise" might be of sufficiently similar meaning, such relaxed usage is a recipe for confusion in quantum physics. ..... Paul Kinsler

        Quite so, PK, however, at certain levels/heights/depths of practical research and virtual endeavour does it provide a very effective stealthy, stay healthy fail-safe security cordon around both free ranging and freelancing assets which be cruising through a vast space of expensive treasury exhausting liabilities.

        The prime bottleneck holdups in research and applications with all things quantum, be they QMechanics or QPhysics or QComputers or QCommunication, are as a result of thinking there are rules which have been followed and are further to be followed for rapid progress into the future, rather than simply realising the opposite is true and the correct path to shared lofty aims and proposed stellar goals. There are no rules to regulate with hinderance/interference and noise/industry. The fields in front of you are pure virgin terrain.

    2. This post has been deleted by its author

  2. Neil Barnes Silver badge

    require materials that haven't yet been developed

    Surely not - there's a box of them on the shelf there, right next to the unobtanium!

  3. MarkET

    photonic cat qubit

    How on earth do you debug a 'maybe' state...

    Yes boss, half of my code may have worked before lunch but I'm not sure now.

    1. Tom 7

      Re: photonic cat qubit

      You keep repeating the calculation until you have enough runs for it to be statistically significant.

      1. MarkET

        Re: photonic cat qubit

        Yes, precisely. Or not.

      2. bombastic bob Silver badge
        Boffin

        Re: photonic cat qubit

        this trick works with analog measurements. I've done it. Sample noisy thing 100 times, and do a median and standard deviation to determine how accurate your measurement is. Moving averages also help when the value is expected to change. In theory, once you have a really good data set, and you throw out anything that is outside of 'n' standard deviations, you should have good results. Maybe.

        Also have been reading up on QM lately, out of curiosity. Found THIS:

        https://www.forbes.com/sites/chadorzel/2017/02/28/how-do-you-create-quantum-entanglement/

        I guess you could say that the main problem is creating entangled particles reliably, and THEN keeping them around long enough to actually be useful.

        also this one: https://erc.europa.eu/projects-figures/stories/how-entangle-two-electrons-%E2%80%93-and-do-it-again-and-again

      3. zuckzuckgo Silver badge

        Re: photonic cat qubit

        Or, if it is easier to check the answer than to find the answer, just keep going to you get the right one.

    2. fidodogbreath

      Re: photonic cat qubit

      Yes boss, half of my code may have worked before lunch but I'm not sure now.

      That seems to be the approach at Microsoft.

    3. fajensen

      Re: photonic cat qubit

      Easy: You create a quantum channel that sends information back in time, from when the outcome is known, to "now".

  4. Jason Bloomberg Silver badge
    Alien

    It's all beyond me

    "It involves creating a cavity with two mirrors and bouncing photons back and forth so the light creates an interference pattern that represents quantum information.

    The shiny monoliths at Utah and elsewhere all seem to be three sided. Perhaps we're being given a clue?

  5. Aristotles slow and dimwitted horse

    That poor cat.

    Its been in that box for so long the poor thing is probably in such a state that it doesn't know whether its coming or going.

    Someone call the RSPCA.

  6. el kabong

    Just keep adding stuff to it and you will end up with a classical contraption not a quantum one

    Basically each additional step you take to make your quantum contraption work properly is a step that makes it more classical, less quantum.

    I guess nature is trying to tell us something, something very important. If only we would listen.

    If were wise I would not put too much faith in those efforts.

  7. Long John Silver
    Pirate

    If the answer is not 42 then ...

    ... you know the quantum computer is faulty.

    1. ClockworkOwl
      WTF?

      Re: If the answer is not 42 then ...

      If all the elements add up to the final result, then you've got your Bistromathics wrong!

  8. jonathan keith
    WTF?

    Top-class boffinry

    ... although they lost me at 'cat state'.

    Alternatively, witchcraft.

  9. Gene Cash Silver badge
    Coat

    cat state

    cat: state: No such file or directory

  10. David Roberts
    Trollface

    photonic cat qubits

    More special than the dog's bollocks.

  11. CrackedNoggin Bronze badge

    "That's why scaling current 50-100 qubit quantum computers up past a million qubits, in order to run practical, general-use quantum systems ..."

    According to [ https://spectrum.ieee.org/tech-talk/computing/hardware/qubit-supremacy ]

    "Assuming these quantum circuits were competing against supercomputers capable of up to a quintillion (1018) floating-point operations per second (FLOPS), the researchers calculated that quantum supremacy could be reached with 208 qubits with IQP circuits, 420 qubits with QAOA circuits and 98 photons with boson sampling circuits."

    A QC with 100 bits already has 2**100 > 10**30 states.

    "To correct arbitrary 1-qubit errors, the technique will reduce required overhead – the extra qubits needed for fixing things – by a factor of three, he said. Lieu said the technology to demonstrate this theory is within reach ..."

    Ladbroke's is using a QC to compute the odds that the amount of extra hardware and/or time required for QC error correction is going to be greater than O(N) for N quantum bits.

    1. diodesign (Written by Reg staff) Silver badge

      A million qubits

      The one-million-plus figure is the ball-park estimate thrown around by IBM, Intel, China, and others, in terms of building something generally useful.

      C.

      1. Michael Wojcik Silver badge

        Re: A million qubits

        Right. Supremacy != generally useful. Also, as Aaronson pointed out recently, supremacy is a slippery beast; it's not trivial to determine where the classical cutoff might be for some of these problem domains.

        Supremacy has lost most of its utility as a measure of QC progress.

        And, of course, even if we eventually have a reliably-working "generally useful" QC machine, that says nothing about the economics. It might be able to solve only a handful of interesting problems in an economically-feasible fashion.

        (Ironically, again according to Aaronson, the recent Pan & Lu BosonSampling experiment was constrained by the economics of classical computing: "A couple weeks later, the authors responded, saying that they’d now verified their results up to n=40, but it burned $400,000 worth of supercomputer time so they decided to stop there." But that's just a reminder that ultimately it's the economics that will dominate "useful".)

  12. Anonymous Coward
    Anonymous Coward

    FTFY

    Useful quantum computers will be impossible without error correction.

    1. Tail Up
      Pint

      Re: FTFY

      and "640K ought to be enough for anybody" (cited from anno 1981 BC)

  13. Torben Mogensen

    "All we have to do is put them together"

    That must be the understatement of the decade. Problems arise in quantum computer exactly when you put elements together. Each element may perform predictably on its own, but when you put them together, chaos ensues.

  14. bigmacbear

    "Erwin, what have you done with the cat? ...

    ... It looks half-dead." -- Frau Schrödinger

    1. chivo243 Silver badge
      Coat

      Re: "Erwin, what have you done with the cat? ...

      What if the cat in the box could come back to life?

      He'd be scratching the lid like hell, trying to get un-buried in the garden...

      What's this atom in my coat pocket?

  15. Tail Up
    Boffin

    Obvious/Incredible

    @ o these never repeating sequences of noise/

    > о, сколько нам открытий чудных

    готовит просвещенья дух

    и опыт, сын ошибок трудных

    и гений, парадоксов друг

    и случай, БОГ- изобретатель...<

    ; А.С. Пушкин

    \eof

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like