Wednesday, February 08, 2017 ... Deutsch/Español/Related posts from blogosphere

Entanglement tests with star-powered pseudorandom generators are childish

Quanta Magazine's Natalie Wolchover has described an experiment performed with the help of many scientists, including some very famous ones, and published in the prestigious PRL

Experiment Reaffirms Quantum Weirdness (Quanta Mag.)

Cosmic Bell Test: Measurement Settings from Milky Way Stars (PRL, PDF full).
As far as I can say, the people who really need to perform similar tests of the laws of Nature – because they're not sure about the outcome – deserve a nice bed in a psychiatric asylum, not pages in the PRL. Gentlemen, this is just stupid beyond imagination. The fact that the insanity of the papers on "quantum foundations" has reached similar proportions is a worrisome testimony to the limitless anti-quantum bigotry of so many people.

OK, what happened?

The correlations between measurements of two subsystems that have interacted in the past (or, more typically, resulted from a decay of one particle), as predicted by quantum mechanics, may be stronger and tighter than the correlations predicted by any local classical (="realist") theory. An insanely overrated example of this trivial fact was discussed by John Bell. Everyone who followed the progress in physics has known since 1925 or 1926 (or from his birth, whichever event came later) that the laws of physics are not classical (="realist").

So even though they are local, as Einstein's special relativity established in 1905, they often give predictions that no classical theory could. For example, Nature is capable of violating Bell's inequalities. Local classical theories of Nature would imply that some increasing function of correlation coefficients in an experiment with two spins is always smaller than or equal to a particular bound; quantum mechanics as well as experiments show that the function of the correlation coefficients is greater. OK, Nature isn't described by a local classical theory. The actual laws of Nature may achieve higher correlation coefficients than any local classical theory. Sometimes, they may imply a lower correlation. In other cases, the laws of quantum mechanics just make totally different predictions than classical theories and the differences may be qualitative from many perspectives.

In the Bell setup, you measure some correlations between polarizations of two particles. For each particle in the pair, you may measure the polarization (one bit of information, e.g. whether the spin is up or down) with respect to a different axis or a different kind of polarization. Let me refer to this choice "what you measure" as the "choice of axes". For different combinations of the two axes (for the two particles in the pair), quantum mechanics predicts some correlation (in the interval from \(-1\) and \(+1\)) between the two outcomes of the experiment (which are equivalent to two bits).

The predictions of quantum mechanics make perfect sense and they agree with the experiments. But some people hate them so much that even 90+ years after quantum mechanics was established, they are still inventing alternative, non-quantum but non-local, theories that could perhaps make the same predictions. No such a theory may ever be viable and I've discussed the reasons many times.

But here, an even more insane "competitor" of the quantum explanation was being proposed – and tested by professional (="paid by institutions for their work") physicists. Maybe the measured correlations between the polarizations are higher than those of a local classical theory because the experimenters who are choosing the axes are not "free". Something invisible controls their brains and in every repetition of the experiment with the two particles, they are selectively choosing such axes that the statistics of the measured outcomes seems to confirm quantum mechanics – and violate any local classical theory – even though if the experimenters were really "free" in their choice of the axes, they would see that Nature actually obeys the laws of a local classical theory.

The degree of attachment of some people to classical physics is just amazing.

OK, in 2010, a man named Michael Hall has argued in a psychiatrically ill paper – also in PRL (!) – that conspiracy theories of this kind are great. You might say that he was an isolated lunatic. But yesterday, PRL published a paper about the experiment (the hyperlink is at the top) performed (or contributed to) by 20 authors. The list of authors includes Anton Zeilinger, a well-known entanglement experimenter from Vienna, and Alan Guth, who recently invited Zeilinger to MIT.

What have they done? They have done the "random choice of the axes" determining the "type of the polarization measurement" differently than before. Before, humans could choose the axis freely or some pseudorandom generator using a nearby computer was used. But the computer could have been a part of the conspiracy, they were worried. The calculation of the pseudorandom numbers on the computer could have been restricted by a great puppet master of the Universe so that the pseudorandom numbers weren't so close to random numbers, after all. Instead, they were chosen so that the resulting measurements of the polarizations indicated that quantum mechanics works even though, in reality, it's local classical physics enhanced by a puppet master that describes the Universe!

This whole "loophole" is just plain insanity – and as far as I can say, it's even logically nonsensical because if the pseudorandom generator by the computer is deterministic, there's just no way to modify its individually spitted numbers separately.

So these 20 people haven't used a computer as the source of the pseudorandom numbers, being worried that the computers could be controlled by the puppet master in the giant conspiracy. Instead, they have pointed their telescopes to some Milky Way stars that are hundreds or thousands of light years away. Their slightly oscillating colors etc. have been translated to numbers and those were used to calculate the pseudorandom numbers.

Now, wait for a few hours to get sober (especially if your name is Tony) and then try to answer the following question: Do you expect statistically different correlations in the experiment in which the pseudorandom numbers are chosen by a computer; or in the experiment where the pseudorandom numbers are chosen by the stars?

Needless to say, it just f*cking doesn't matter how you choose your pseudorandom numbers. Obviously, that's the conclusion of their experiment, too. For all practical and experimentally detectable purposes, the axes are being chosen randomly – whether you choose a purely computer-based pseudorandom generator or whether you use some random numbers from the stars as the input. Neither the computer nor the stars of the Milky Way have the "purpose" of producing qualitatively misleading results of some petty experiments. They are not parts of a giant cosmic conspiracy. And even if there were a giant cosmic conspiracy, the top conspirators' goal would surely not be to make the outcome of some irrelevant experiments misleading.

But all these people seem to pretend that they believe that what they have done is serious science. Sorry, it's not. Wolchover calls the ultimate conspiracy theory "an intriguing loophole" and in a tweet, she wrote that it was "sad" that this obviously ludicrous possibility was booted. And Guth, Zeilinger, and pals are seriously deducing constraints from their experiment – whose result had to be obvious to every sane person. For example, they tell us that
if there is a cosmic conspiracy that determines which axes you may want to choose, the conspiracy must also involve and constrain stars that emitted the relevant starlight before William Shakespeare wrote a sentence about uncertainty.
The Shakespeare-themed translation of "at least 600 years in the past" was invented by Michael Hall, the lunatic who wrote the original 2010 paper.

I won't discuss some technicalities in their paper seriously because I refuse to pretend that papers like that are serious. You know, the problem is that this community is increasingly detached from the reality and their experiments. Every experiment has said – and almost certainly will say – that quantum mechanics works while any empirically inequivalent proposed alternative is demonstrably wrong, usually brutally so.

If they actually cared about the experiments, they would learn one lesson: their confidence in quantum mechanics should increase and they should spend less time with similar alternatives to quantum mechanics and similar efforts to question quantum mechanics because these alternatives and efforts are wrong, stupid, and increasingly so. However, these people – as a community and as individuals – draw exactly the opposite lesson. They are proposing and testing alternatives to quantum mechanics that are increasingly far from what the experiments actually tell you to think.

These people aren't doing science because while they are making experiments, they don't give a damn about their results. The results never affect their reasoning and their choice of subsequent experiments.

Moreover: If you're insane enough and willing to consider the "hypothesis" that the computer doesn't have the "freedom" to spit basically arbitrary pseudorandom numbers, you will surely be willing to assume that Shakespeare and the stars when they were emitting their light 600 years ago don't have this freedom, either. They won't care that a 600-year-old starlight was used to calculate the pseudorandom numbers because the stars may be puppets in the conspiracy, too. Everything that anyone has ever observed or will ever observe may give misleading results that result from a cosmic conspiracy that restricts everyone's and everything's freedom according to some strangely correlated patterns.

We can't ever rigorously disprove the most far-reaching types of these conspiracy theories. But that doesn't mean that they're scientific. They're not scientific because they're not making any predictions that would be analogous to the well-defined predictions of quantum mechanical theories. Let me emphasize that the criticism isn't that these alternative theories say "just as much as" quantum mechanics does while they need to say more. If that were true, it would be fine and they could be viewed as serious alternatives (that's the situation of string theory which surely makes basically the same predictions for low-energy experiments as quantum field theories and some "critics" incorrectly say that it's not enough to consider string theory). The problem is that these theories don't make any of the major valid predictions that quantum mechanics does. They don't say anything that resembles the predictions done by science.

Their defenders may say:
The world only looks local, Lorentz-symmetric, quantum, Darwin-evolving... even though in reality it is non-local, Lorentz-violating, classical, creationist... (choose your favorite adjectives). The illusion that the world is local, Lorentz-symmetric, quantum, Darwin-evolving etc. is caused by a giant cosmic conspiracy whose purpose is to only allow the choice of axis that makes the laws of physics look completely different than they actually are – and they actually are non-local, Lorentz-violating, classical, creationist, full of tooth fairies and the Papa Smurf flying on an Iron Men etc.
But you're not giving any evidence whatsoever for any of the unusual alternative facts or proposed properties of Nature (such as those about the Papa Smurf). Your choice of these adjectives and properties reflects nothing else than your prejudices and the flavor of your mental disease. Instead, whenever you make a test, you find out that stupid ideas are stupid, indeed. Is there a limit beyond which the people seriously considering these alternative explanations should be classified as mentally ill? As far as I am concerned, you have already transgressed the red line by some 600 light years, Alan et al.

Science – and any lesson we may learn from it – depends on the assumption that the experimenters have the underlying freedom to measure what they want and that their decisions are independent and may be assumed to be independent from particular distant events. If some cosmic conspiracy tells us that "we are not allowed to want to choose the axis in one way or another at a given moment", well, it's fine, and our observations might give us a misleading testimony about the actual properties of Nature. The laws that were extracted from our observations – local, Lorentz-invariant, quantum mechanical laws with Darwin's evolution etc. – "only" tell us what the world looks like after the restrictions from the cosmic conspiracy are imposed.

But you know what? This "only" is demagogic and the words "looks like" don't really diminish the results of science at all. At the end, the purpose of science is to tell us as accurately as possible what the Universe will look like to us. So if it looks like it is local and quantum mechanical in all our experiments, the scientific understanding of the truth implies that the Universe indeed is local and quantum mechanical, and it's the alternatives that are simply wrong, regardless of your emotional attachment to these insanities that you may develop.

Because of its fundamental reliance on the empirical data, natural science defines the truth about Nature as the propositions that follow from the most careful evaluation of the most accurate actual experiments that may be done. So if the cosmic puppet masters and conspirators do their job so well that Nature looks local and quantum mechanical in all of experiments that are in principle doable, then the cosmic puppet masters have indeed succeeded in changing the truth itself: They made the Universe local and quantum mechanical according to science!

I wouldn't allow the PRL paper to be published on this blog because it wouldn't surpass the elementary quality threshold that is required from guest bloggers. The standards of PRL are sometimes way lower than those of TRF.

By the way, Wolchover's article contains lots of the usual misconceptions about quantum mechanics and entanglement. For example, she wrote:
Measure photon A to be vertically polarized, and photon B instantaneously becomes horizontally polarized, even though B’s state was unspecified a moment earlier and no signal has had time to travel between them. This is the “spooky action” that Einstein was famously skeptical about...
This is a misleading formulation that indicates that there is a "spooky action". But the actual experiment of this kind contains no evidence for the action at a distance and our established theories unequivocally say that there is no action at a distance. The corrected formulation would say:
Measure photon A and find it to be vertically polarized. In this way, because of the entanglement (quantum mechanical description of the pre-existing correlations), you also immediately learn that the photon B is horizontally polarized, something you didn't know a moment earlier. No signal needed to travel between them. Einstein incorrectly assumed that the polarizations must have had some objective values that are in principle knowable even before the measurements, and all candidate theories obeying this (wrong) Einstein's condition would need an actual signal to propagate in between the two places, a problem that Einstein called a "spooky action at a distance" which would contradict his special relativity. But Nature doesn't obey Einstein's assumption that the polarizations are in principle knowable even before the experiments and that's what allows quantum mechanics to guarantee the correlations even in the absence of any signal.
And there are several other pieces of the fake news in her article that I have debunked about 1,000 times but science writers are simply incapable of learning these basics of modern physics properly so it's a waste of time to try to teach them.

Add to Digg this Add to reddit

snail feedback (0) :

(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){ (i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o), m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m) })(window,document,'script','//','ga'); ga('create', 'UA-1828728-1', 'auto'); ga('send', 'pageview');