Well worth reading and trying to understand Re Quantum computers will never overcome noise issues?

["Dave Farber" <dave () farber net>]

---------- Forwarded message ---------
From: Rodney Van Meter <rdv () sfc wide ad jp>
Date: Thu, Feb 15, 2018 at 7:17 PM
Subject: Re: [IP] Re Quantum computers will never overcome noise issues?
To: David Farber <dave () farber net>
CC: Rodney Van Meter <rdv () sfc wide ad jp>, ip <ip () listbox com>


Dave,

Apologies for the slow reply; I put this on the stack to answer, and just
now got around to it.

Kalai first, then Joe…

Gil Kalai is a well-known quantum computing skeptic. He is a serious
researcher, and deserves to be taken neither more nor less seriously than a
contrarian in any field. Many of the most famous quantum computing
theorists, most especially Aram Harrow but including Scott Aaronson, Dave
Bacon, and Peter Shor have engaged with him.

Fundamentally, Kalai doesn’t believe that quantum error correction will
work. He believes that errors in quantum systems will be (classically)
correlated in a way that will destroy quantum states. He’s asking an
important question there; correlated errors was the first thing that really
worried me about quantum error correction, as well. I finally worked
through it to my satisfaction. However, Kalai is more sophisticated than I
am mathematically, and he is engaging with the question at a deeper level
than I did.

But fundamentally the models that we have are known to correspond to what
we know about noise in quantum systems, and if those conditions hold, then
QEC will work. Kalai does not dispute this; what he disputes is whether the
noise models cover what will really happen. He’s not saying we’re applying
our model wrong, he’s saying we might have the wrong model.

Given that no one has yet built a system with a large number of entangled
quantum sub-systems, it could be that he is right. But that would mean NEW
PHYSICS, not just new equations. It’s hard to imagine that anyone would be
more excited than the theorists I mentioned above if that proved to be the
case. It would be a profound discovery. (See Scott Aaronson’s Ph.D. thesis
for a discussion of such an eventuality.
https://www.scottaaronson.com/thesis.html)

Kalai’s work is all highly speculative, a solution in search of a problem.
There are no experimental hints of anything unexplained that would require
such new physics. It doesn’t really derive from something fundamental, it’s
more of, “In order for quantum computers to be unbuildable, an effect that
looks like this would have to exist.” So how it would most likely play out
is a decade of increasingly precise quantum experiments, experimentalists
continually frustrated at their inability to build, manipulate and measure
large-scale entanglement. Noise source and inaccuracy after noise source
and inaccuracy are eliminated. Eventually, someone recognizes a pattern in
the behavior, realizes that it corresponds to some equation, writes down
that equation. Later, someone recognizes that it corresponds to something
that Kalai wrote down in the early 2010s. The phenomenon is given a name,
and a search for a physical mechanism that would generate the behavior
begins. Nobel Prizes all around.

A second option is that we perpetually remain unable to reduce noise to
that level, leaving us in an ambigous state, unable to prove that quantum
computing works but with no reason to believe it doesn’t.

Far more likely, in my opinion, is that quantum error correction works
exactly as advertised, and we will eventually reduce noise enough to do
truly fault-tolerant computation. In fact, it’s likely that within about
three years we will see exactly that; it could happen as early as this year.

—

Joe (another well-known contrarian) is on the right track. It has been
known since the 90s, I think, that independent quantum computers without
entanglement between them can’t cooperate to solve a larger problem more
efficiently. Yepez referred to this as “Type-II” quantum computers (
https://pdfs.semanticscholar.org/e653/6326e6e31503e348ce005fa049f3532fba83.pdf).
So, interconnecting small quantum computers to make a larger, entangled
quantum multicomputer is one of the most urgent work items on any
experimentalist’s desk. Ion trappers (e.g., Chris Monroe and Jungsang Kim,
who have created IonQ, a startup dedicated to exactly this) are farthest
along in this kind of architecture. (My thesis was on this very topic,
https://arxiv.org/abs/quant-ph/0607065)

This is a particular Achilles heel for transmon (the most advanced form of
superconducting) systems, which otherwise are getting the most attention
and investment these days. Unfortunately, due to the low energy levels they
work at, they most naturally emit microwave photons, which are extremely
hard to capture and manipulate one-by-one. So, there are several proposals
on the table for how to convert a transmon state to an optical or infrared
photon. But, just recently, Andreas Wallraff’s group (apparently) completed
the amazing feat of coupling two transmons over 80cm of coax
(https://arxiv.org/abs/1712.08593). All of this bodes well for scalability.
Control is still a major issue for scalability, though.

Enough for now.

FYI, the next run of our MOOC, Understanding Quantum Computers, is
tentatively scheduled for mid-April. We are adding new material, revising
some based on feedback and recording interviews with several additional
researchers to add depth. Look for it!
https://www.futurelearn.com/courses/intro-to-quantum-computing/

—Rod


On Feb 10, 2018, at 2:18 AM, Dave Farber <dave () farber net> wrote:


---------- Forwarded message ---------
From: Joe Touch <touch () strayalpha com>
Date: Fri, Feb 9, 2018 at 9:13 AM
Subject: Re: [IP] Quantum computers will never overcome noise issues?
To: David Jack Farber <dave () farber net>
CC: ip <ip () listbox com>


Hi, Dave,

IMO, the biggest impediment is scale.

There have been many quantum computers announced with increasing numbers of
qubits - but those announcements are misleading, IMO.

An 8-qubit computer solves 8-bit problems instantly only when all 8 qubits
are completely entangled with each other. Large quantum computers typically
involve many small quantum computers interconnected classically - and those
interconnections destroy the quantum coherence and entanglement that gives
quantum computing its edge.

Not that I disagree with the noise issue, but even if that’s solved,
there’s a scale issue too.

Joe

On Feb 8, 2018, at 5:06 PM, Dave Farber <dave () farber net> wrote:


---------- Forwarded message ---------
From: Allen <allenpmd () gmail com>
Date: Thu, Feb 8, 2018 at 1:28 PM
Subject: [Cryptography] Quantum computers will never overcome noise issues?
To: Crypto <cryptography () metzdowd com>


Gil Kalai, a mathematician at Hebrew University, believes quantum
computers will never be able to overcome noise issues. See article at:

https://www.quantamagazine.org/gil-kalais-argument-against-quantum-computers-20180207/

Any thoughts on this?
_______________________________________________
The cryptography mailing list
cryptography () metzdowd com
http://www.metzdowd.com/mailman/listinfo/cryptography



This message was sent to the list address and trashed, but can be found
online.
<https://www.listbox.com/login/messages/view/20180215191741:CBAAEB02-12AE-11E8-B415-CE7F7669CA45/>



-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/18849915-ae8fa580
Modify Your Subscription: https://www.listbox.com/member/?member_id=18849915&id_secret=18849915-aa268125
Unsubscribe Now: 
https://www.listbox.com/unsubscribe/?member_id=18849915&id_secret=18849915-32545cb4&post_id=20180215192329:9B047A30-12AF-11E8-8E5B-8141216728AB
Powered by Listbox: http://www.listbox.com