Worth noting that this document was (for whatever reason removed) after initial publication:
* https://ntrs.nasa.gov/search.jsp?R=20190030475
* https://fortune.com/2019/09/20/google-claims-quantum-suprema...
Quantum Supremacy Using a Programmable Superconducting Processor
51–60 of 71 posts
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#52Here's a Twitter thread that sounds like this is a very unusual take on "quantum supremacy":
https://twitter.com/isislovecruft/status/1175557089067823104
I'm not deep enough into the topic to judge. Thoughts?
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#53These seem to be much stronger claims than were made last time this was discussed: ”””While our processor takes about 200 seconds to sample one instance of the quantum circuit 1 million times, a state-of-the-art supercomputer would require approximately 10,000 years to perform the equivalent task.””” ”””Quantum processors based on superconducting qubits can now perform computations in a Hilbert space of dimension 2^5…
Yes, quantum supremacy disproves the extended church-turing thesis (Bernstein-Vazirani). This thesis states that any computation that can be computed efficiently can be computed efficiently with a classical computer (ie a Turing machine). ECT thesis means that quantum computers will never be able to compute something significantly faster that classical ones. However, technically, google haven't proved yet quantum sup…
> However, technically, google haven't proved yet quantum supremacy
To be absolutely clear, an experiment (or a set of replicated experiments) can't prove quantum supremacy. Quantum supremacy is a statement about about the asymptotic behavior of quantum and classical computers, and obviously, a finite set of data points cannot prove anything about asymptotes.
What these experiments are doing are gathering evidence against the extended Church-Turing thesis - in the same way other experiments are used to gather evidence in favor of/against general relativity or some other physical theory.
The proof of quantum supremacy within the framework of the currently accepted laws of physics (quantum theory), has to be a mathematical proof. Currently, we have no idea how to construct such a proof.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#54Earlier quoted context omitted.
There is a slight difference, that really matters to theoretical computer scientists (as in complexity theory) and physicists working on quantum computing: The challenge is to construct a physical experiment that consists of only 50ish "elementary" and "programmable" building blocks that is infeasible to simulate on a classical computer. This "universality" and small number of elements is what is exciting and what ma…
I get that its cool that it’s quantum and that the circuits are programmable, sure. My point was that the comparison to the classical computer makes little sense.
You can read Feynman's original paper about quantum computers, which is easy to read and makes clear why programmability is important https://catonmat.net/ftp/simulating-physics-with-computers-r...
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#55Earlier quoted context omitted.
No, "quantum supremacy" is a technical term that has not been achieved before, and has not been claimed to be achieved by anyone serious. You may be misremembering previous claims called something different.
I guess you're right, I remember seeing quantum algorithms that were assumed to be more efficient than any classical algorithm and then people devised equally efficient classical algorithms. E.g. work by Ewin Tang https://arxiv.org/abs/1811.00414
They were definitely not being assumed to be faster. It was considered very uncertain. Quantum algorithms for which people are pretty confident there is no classical algorithm of similar speed, while still unproven, are Shor's algorithm (i.e., no fast classical factoring) and the present case of sampling for supremacy.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#56Worth noting that this document was (for whatever reason removed) after initial publication: * https://ntrs.nasa.gov/search.jsp?R=20190030475 * https://fortune.com/2019/09/20/google-claims-quantum-suprema...
I would put on my tinfoil hat and call for a conspiracy theory, but the reality is probably a lot more boring.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#57"However, realizing the full promise of quantum computing (e.g. Shor’s algorithm for factoring) still requires technical leaps to engineer fault-tolerant logical qubits" This is NOT how the media has chosen to report on this. It's always quite frustrating to figure out how a paper diverges from the press it gets.
Well I mean it's still pretty big though. We know for a fact now that there are only technicalities like fault tolerance in our way.
Still you're right though, the race for "fist post" by the media definitely exaggerates scientific breakthroughs. To the science specific literate they can sift through the hype but others get it in their head it's something else. Like investors for example, those guys are big dreamers and when they hear stuff like "ground breaking quantum supremacy" they'll throw everything they got at it. Investing in startups by itself has worse odds than slot machines, hype only makes it harder to predict the future. But oh well, at least we have the actual details and most of the articles link the results for those actually interested. Can't do much about what others think or do though.
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#58Here's a Twitter thread that sounds like this is a very unusual take on "quantum supremacy": https://twitter.com/isislovecruft/status/1175557089067823104 I'm not deep enough into the topic to judge. Thoughts?
Here's the relevant part of the twitter rant- seems interesting but I can't tell myself whether it makes sense or not :)
"they repeatedly asked the quantum processor for outputs to a random circuit, and the “quantum supreme computation” was that the outputs were not actually random, but subject to bias due to quantum interference, which they check by solving the path integral in classical simulation
which, to me, is an infuriating thing to label “quantum supremacy” because quantum interference is EXACTLY WHY WE CAN’T USE CURRENT QUANTUM PROCESSORS TO “BREAK ALL CRYPTO” IN THE FIRST PLACE BECAUSE WE CANNOT DO EFFICIENT QUANTUM ERROR CORRECTING CODES
it’s like pointing at a fucking Zener diode and saying, “quantum supremacy” simply because quantum electromagnetic domain wall tunnelling happens at random, followed by a non-justification of “well your iphone can’t do that”"
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#59Here's a Twitter thread that sounds like this is a very unusual take on "quantum supremacy": https://twitter.com/isislovecruft/status/1175557089067823104 I'm not deep enough into the topic to judge. Thoughts?
Here's the relevant part of the twitter rant- seems interesting but I can't tell myself whether it makes sense or not :) "they repeatedly asked the quantum processor for outputs to a random circuit, and the “quantum supreme computation” was that the outputs were not actually random, but subject to bias due to quantum interference, which they check by solving the path integral in classical simulation which, to me, is…
Sorry but no. The poster is taking a simplistic view of things. Random circuit sampling is a generally accepted method of establishing quantum supremacy.
See this paper from Dr. Vazirani (Berkeley) saying as much https://arxiv.org/pdf/1803.04402.pdf
Re: Quantum Supremacy Using a Programmable Superconducting Processor
#60To me this reads a lot like “quantum computer faster than a classical computer at ‘simulating’ itself”. Constructing a physical experiment where it would take a classical computer 10000+ years to simulate the outcome is trivial. I don’t see how this is very different.
This is not a one-off. This is a general purpose computer with general purpose gates. Verifying the correctness of the RCS is nontrivial.