minkwe wrote:
... I dare say "fraud" has not been ruled out.
Wow!
***
minkwe wrote:
... I dare say "fraud" has not been ruled out.
Joy Christian wrote:minkwe wrote:... I dare say "fraud" has not been ruled out.
Wow!
***
minkwe wrote:Perhaps you are hinting at the question of how the data came to be. And that is an wide open question, and though I'm not making any suggestion here, I dare say "fraud" has not been ruled out.
localyokel wrote:Are you suggesting they are concealing data, that is that they have gone ahead and hidden variables, that when analyzed would prove the existence of hidden variables?
gill1109 wrote:There is no "post-selection" in the Hensen et al. experiment. There is something you might call "pre-selection". It does not constitute a "loophole".
It is a three-party Bell-type experiment. There are three locations: let's call them the labs of Alice, Bob and Caspar. An electron spin in a Nitrogen-Vacancy defect in a diamond crystal is repeatedly excited by a laser in both Alice's and Bob's laboratories. Occasionally, both spins emit a photon which occasionally, both are detected on the other side of a birefringent polarizing beam splitter in Casper's lab, where there are two photo-detectors. Occasionally, both of those detectors click. If so, one doesn't know which photon went which way. On those occasions, according to conventional QM, Alice and Bob's two spins were entangled during a small time interval, during which also the two spins rapidly got measured in randomly chosen directions.
All the time, random settings are being chosen in Alice and Bob's labs, the spins of the two electrons in the NV defects are being measured, the settings and outcomes are registered, the spins are "reset" and then excited again.
The experimenters are interested in the correlations between Alice and Bob's outcomes, given Alice and Bob's settings, and given Casper's two detections.
There is no direct point in keeping hold of all the data which Alice and Bob temporarily had in their possession on the occasions when Caspar did not get two clicks. You can read in the original papers what was the thinking here and what happened to it. It no longer exists. Crying fraud and demanding to see it suggests to me that you didn't actually take much notice of the published description of the experiment.
If the scientists of the world think it is important, they had better get the taxpayers of the world to fund a new experiment. As I said before, it would be smart to have the new experiment performed by different research groups in different places, and while they are at it, to run the experiment for about 10 times as long, so as, hopefully, to get a five sigma deviation from the predictions of local realism, which is what one can predict, assuming that the true value of S was indeed, as QM theory predicts, and the data certainly suggests, about 2.4 (the data gives this measured parameter a standard deviation of about 0.2). That's a value about halfway between the LR bound of 2 and the QM bound of 2.828...
As Fred says, the results of the Delft experiment fit perfectly well to what quantum mechanics predicts, there is no need whatever to suspect any kind of fraud. It is however true that the size of the experiment is much too small to have had much chance of detecting the opposite (any departure from QM).
I haven't seen Joy Christian's specific modelling of this particular three-party Bell experiment, nor how his theory should be extended or adapted to deal with observed conditional correlations between Alice and Bob's outcomes, conditional on Alice's, Bob's and Caspar's settings and Caspar's outcomes. As I understand his writings, he argues that it is un-physical to compare one correlation to three others, each of the four measured on data collected in distinct sub-experiments. But he happily himself derives the Tsirelson bound of 2 sqrt 2 and the no-signalling bound of 4. Fred also confirms that the no-signalling bound of 4 is a valid bound. Fred and Joy claim that their data-rejection simulations can be set to illustrate that all three bounds (2, 2 sqrt 2, and 4) can be achieved in appropriate circumstances. So I don't understand the remarks that experimenters "used the wrong inequality" since "you can't violate a mathematical inequality" anyway. Obviously, no-one is going to get the Nobel prize, nor the Abel prize if they are young enough, for showing that if you have four numbers between +/-1, then the sum of three of them minus the fourth can't exceed 4.
The so-called "no-signalling" bound (which could be attained by so-called Popescu-Rohrlich boxes if they existed) is rather trivial in this case. The principle of "information causality" introduced by Marcin Pawlowski and his collaborators a few years ago shows that the Tsirelson bound of 2 sqrt 2, which follows from assuming validity of QM (but not necessarily two dimensional quantum systems and perfect spin measurements), shows that the 2 sqrt 2 bound follows from general causality principles which intuitively should hold even if QM is not true, i.e., the principles should hold, and hence the bound be true, in much greater universality in reasonable physical theories "beyond QM".
gill1109 wrote: It no longer exists.
local wrote:gill1109 wrote: It no longer exists.
It's stunning that Gill now so calmly acknowledges that nobody will ever be able to see the data to independently analyze the Delft (Hensen et al) experiment, after reassuring us he would get it from his friends. Dog ate the homework! The Nobel committee will not be amused. Hensen et al are unable to disprove the hypothesis that they manipulated the data to produce only an artifactual violation. They ask us to trust them.
Let's be crystal clear. Gill informs us portions of the data were selectively destroyed. I claim that this destruction was intentionally biased to produce only an artifactual violation.
Ancel Keys would be proud.
gill1109 wrote:There is no "post-selection" in the Hensen et al. experiment. There is something you might call "pre-selection". It does not constitute a "loophole".
It is a three-party Bell-type experiment. There are three locations: let's call them the labs of Alice, Bob and Caspar. An electron spin in a Nitrogen-Vacancy defect in a diamond crystal is repeatedly excited by a laser in both Alice's and Bob's laboratories. Occasionally, both spins emit a photon which occasionally, both are detected on the other side of a birefringent polarizing beam splitter in Casper's lab, where there are two photo-detectors. Occasionally, both of those detectors click. If so, one doesn't know which photon went which way. On those occasions, according to conventional QM, Alice and Bob's two spins were entangled during a small time interval, during which also the two spins rapidly got measured in randomly chosen directions.
All the time, random settings are being chosen in Alice and Bob's labs, the spins of the two electrons in the NV defects are being measured, the settings and outcomes are registered, the spins are "reset" and then excited again.
The experimenters are interested in the correlations between Alice and Bob's outcomes, given Alice and Bob's settings, and given Casper's two detections.
There is no direct point in keeping hold of all the data which Alice and Bob temporarily had in their possession on the occasions when Caspar did not get two clicks. You can read in the original papers what was the thinking here and what happened to it. It no longer exists. Crying fraud and demanding to see it suggests to me that you didn't actually take much notice of the published description of the experiment.
If the scientists of the world think it is important, they had better get the taxpayers of the world to fund a new experiment. As I said before, it would be smart to have the new experiment performed by different research groups in different places, and while they are at it, to run the experiment for about 10 times as long, so as, hopefully, to get a five sigma deviation from the predictions of local realism, which is what one can predict, assuming that the true value of S was indeed, as QM theory predicts, and the data certainly suggests, about 2.4 (the data gives this measured parameter a standard deviation of about 0.2). That's a value about halfway between the LR bound of 2 and the QM bound of 2.828...
As Fred says, the results of the Delft experiment fit perfectly well to what quantum mechanics predicts, there is no need whatever to suspect any kind of fraud. It is however true that the size of the experiment is much too small to have had much chance of detecting the opposite (any departure from QM).
I haven't seen Joy Christian's specific modelling of this particular three-party Bell experiment, nor how his theory should be extended or adapted to deal with observed conditional correlations between Alice and Bob's outcomes, conditional on Alice's, Bob's and Caspar's settings and Caspar's outcomes. As I understand his writings, he argues that it is un-physical to compare one correlation to three others, each of the four measured on data collected in distinct sub-experiments. But he happily himself derives the Tsirelson bound of 2 sqrt 2 and the no-signalling bound of 4. Fred also confirms that the no-signalling bound of 4 is a valid bound. Fred and Joy claim that their data-rejection simulations can be set to illustrate that all three bounds (2, 2 sqrt 2, and 4) can be achieved in appropriate circumstances. So I don't understand the remarks that experimenters "used the wrong inequality" since "you can't violate a mathematical inequality" anyway. Obviously, no-one is going to get the Nobel prize, nor the Abel prize if they are young enough, for showing that if you have four numbers between +/-1, then the sum of three of them minus the fourth can't exceed 4.
The so-called "no-signalling" bound (which could be attained by so-called Popescu-Rohrlich boxes if they existed) is rather trivial in this case. The principle of "information causality" introduced by Marcin Pawlowski and his collaborators a few years ago shows that the Tsirelson bound of 2 sqrt 2, which follows from assuming validity of QM (but not necessarily two dimensional quantum systems and perfect spin measurements), shows that the 2 sqrt 2 bound follows from general causality principles which intuitively should hold even if QM is not true, i.e., the principles should hold, and hence the bound be true, in much greater universality in reasonable physical theories "beyond QM".
gill1109 wrote:There is no "post-selection" in the Hensen et al. experiment. There is something you might call "pre-selection". It does not constitute a "loophole".
Hensen et al wrote:Every few hundred milliseconds, the recorded events are transferred
to the PC. During the experiment, about 2 megabyte of data is generated every second. To keep the size of the generated data-set manageable, blocks of about 100000 events are saved to the hard drive only if an entanglement heralding event (E) is present in that block.
minkwe wrote:Why did they record it if not to use it in one way or another, and why would they destroy it (if claims that the data no longer exists is to be believed).
I generate 9 MB of data every 0.1 second for extended periods and transfer large resulting data-sets across the globe on a routine basis
local wrote:Outstanding post, minkwe, bravo!minkwe wrote:Why did they record it if not to use it in one way or another, and why would they destroy it (if claims that the data no longer exists is to be believed).
I can help you with that. I hypothesize that the unmanipulated full data failed to confirm quantum nonlocality and instead confirmed locality (just as we see for Christensen et al). After all the expense and effort by Hensen et al in pursuit of fame, fortunes, and prizes, they could not abide this terrible failure. So they fiddled the data and then destroyed the original full data so that their malfeasance could never be proven, and so that there would not be an experiment on record disconfirming quantum nonlocality. They didn't want to make the same "mistake" as Christensen et al made, i.e., publishing the full data so researchers could analyze it independently.
Joy Christian wrote: If true, then that would certainly qualify to be called "fraud."
In fact, your hypothesis is not an outlier from the historical precedents in science in general or in physics: https://www.amazon.co.uk/Golem-Second-S ... M.+Collins
localyokel wrote:I fear my questions above have been forgotten. Can the group answer:
1. Why should we expect Alice and Bob's transmissions to Casper ever get "entangled"?
2. Is it all local to Casper at the end?
In regards to 2:
Richard says "...Casper's lab, where there are two photo-detectors. Occasionally, both of those detectors click. If so, one doesn't know which photon went which way." So I get the impression that Casper does NOT make two measurements that are separated enough from one another to be called nonlocal.
If the answer to 2 is "it is all local to Casper", then I will put my newly printed out copy of Hensen et. al. through the shredder, and seek other data to decide locality vs. nonlocality.
localyokel wrote:I fear my questions above have been forgotten. Can the group answer:
1. Why should we expect Alice and Bob's transmissions to Casper ever get "entangled"?
2. Is it all local to Casper at the end?
In regards to 2:
Richard says "...Casper's lab, where there are two photo-detectors. Occasionally, both of those detectors click. If so, one doesn't know which photon went which way." So I get the impression that Casper does NOT make two measurements that are separated enough from one another to be called nonlocal.
If the answer to 2 is "it is all local to Casper", then I will put my newly printed out copy of Hensen et. al. through the shredder, and seek other data to decide locality vs. nonlocality.
Return to Sci.Physics.Foundations
Users browsing this forum: No registered users and 83 guests