The (still) open one-sided bet.

Foundations of physics and/or philosophy of physics, and in particular, posts on unresolved or controversial issues

Re: The (still) open one-sided bet.

Postby Joy Christian » Sun May 04, 2014 9:11 am

Heinera wrote:
Joy Christian wrote:I have published a 2D surface simulation of my 3-sphere model for the EPR-Bohm correlation: http://rpubs.com/jjc/16567.

Sigh. Back to Pearle's model then. This looks more and more like an infinite loop.


This is not Pearle's model.

I know you are slow, but your refusal to learn new things is getting tiring: http://libertesphilosophica.info/blog/w ... mplete.pdf
Joy Christian
Research Physicist
 
Posts: 2370
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: The (still) open one-sided bet.

Postby Heinera » Sun May 04, 2014 9:21 am

Joy Christian wrote:
Heinera wrote:
Joy Christian wrote:I have published a 2D surface simulation of my 3-sphere model for the EPR-Bohm correlation: http://rpubs.com/jjc/16567.

Sigh. Back to Pearle's model then. This looks more and more like an infinite loop.


This is not Pearle's model.

Then at least you should get rid of the obvious clues:
Code: Select all
f <- -1 + (2/sqrt(1 + ((3 * eta)/pi)))  ## Pearle's 'r' is arc cosine of 'f'

Jeezus, your new program is nothing but Richard's code with some graphs added.

Use this program to produce the requested file with a list of vectors, then we'll see.
Last edited by Heinera on Sun May 04, 2014 9:40 am, edited 1 time in total.
Heinera
 
Posts: 767
Joined: Thu Feb 06, 2014 1:50 am

Re: The (still) open one-sided bet.

Postby gill1109 » Sun May 04, 2014 9:38 am

Yes this is a simulation of Pearle's model and mainly my code, but with most comments giving some kind of attribution to all the people who contributed code, algorithms, or ideas, removed. Christian does thank Zen aka Chán Satori for the code to draw the 3d surface. I would also like to see little balls floating on the surface at the four points corresponding to the four CHSH setting pairs.

Joy thinks it is a simulation of his model, but whether that is true or not, it *is* a simulation of Pearle's model too. That fact cannot be denied. Not only is it a simulation of Pearle's model, the formulas which Christian has copy-pasted from my simulation of Pearle's model were deduced by me during a week of blood sweat and tears, from Pearle's paper; during which my derivation was also double checked by Florin Moldoveanu. (Not by Christian).

He is learning R by imitation, which is fine, but he also needs to learn good coding (and scientific) practices, e.g. proper attribution of all your sources. He obviously does not ever want to be accused of plagiarism.

Posting the announcement of this Rpubs document here is really off topic in this thread, but OK, it's nice to see Chán's pictures confirming Pearle's and my mathematics.
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby Joy Christian » Sun May 04, 2014 9:45 am

Heinera wrote:Then at least you should get rid of the obvious clues:
Code: Select all
f <- -1 + (2/sqrt(1 + ((3 * eta)/pi)))  ## Pearle's 'r' is arc cosine of 'f'


There is nothing wrong with this line. It is not a clue, but credit to Pearl, for having found the exact analytical solution of the problem. As a supposed mathematician you should know that the same mathematics can be used to describe many different physical phenomena. There is no need to reinvent the wheel if it already exists.
Joy Christian
Research Physicist
 
Posts: 2370
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: The (still) open one-sided bet.

Postby Joy Christian » Sun May 04, 2014 9:51 am

gill1109 wrote:Yes this is a simulation of Pearle's model and mainly my code, but with most comments giving some kind of attribution to all the people who contributed code, algorithms, or ideas, removed. Christian does thank Zen aka Chán Satori for the code to draw the 3d surface. I would also like to see little balls floating on the surface at the four points corresponding to the four CHSH setting pairs.

Joy thinks it is a simulation of his model, but whether that is true or not, it *is* a simulation of Pearle's model too. That fact cannot be denied. Not only is it a simulation of Pearle's model, the formulas which Christian has copy-pasted from my simulation of Pearle's model were deduced by me during a week of blood sweat and tears, from Pearle's paper; during which my derivation was also double checked by Florin Moldoveanu. (Not by Christian).

He is learning R by imitation, which is fine, but he also needs to learn good coding (and scientific) practices, e.g. proper attribution of all your sources. He obviously does not ever want to be accused of plagiarism.

Posting the announcement of this Rpubs document here is really off topic in this thread, but OK, it's nice to see Chán's pictures confirming Pearle's and my mathematics.


Proper credits are already given in the companion simulation: http://rpubs.com/jjc/13965:

## This is S^2 version (aka 3D version: think of S^2 as a surface in R^3)
## This version has been adapted from Richard Gill's optimized version of
## Michel Fodje's original simulation of the model, which can be found here:
## http://rpubs.com/gill1109/EPRB3opt. Later Richard Gill improved his 3D
## version by employing the exact probability distribution derived by Philip
## Pearle in his classic 1970 paper: http://rpubs.com/gill1109/Pearle. It
## should be noted, however, that, unlike Pearle's model, the 3-sphere model
## has nothing whatsoever to do with data rejection or detection loophole.

## All of the above simulations are inspired by the original simulation of
## the 3-sphere model by Chantal Roth, https://github.com/chenopodium/JCS2.

It is also important to note that the simulation I have posted is NOT Pearle's model. To understand why it is not Pearl's model you have to understand the detailed analytical derivations of his model as well as my model, and then the connection between the two models given in this paper: http://libertesphilosophica.info/blog/w ... 1/EPRB.pdf.

The two people who deserve the most credit for ALL the simulations are actually Chantal Roth and Michel Fodje. Without their pioneering attempts to simulate my model, with months of dedicated work in isolation, in entirely civilized and pleasant interaction, is what brought about the present flurry of activities around my work. For years many people, including myself, thought that it was not possible to simulate my model. But my months of intense efforts in collaboration with Chantal Roth changed all that. Then came the breakthrough contribution by Michel Fodje. This was then improved upon by Richard Gill, as I mention in the above credits.

In any case, I have updated the current simulation to reiterate the proper credits from the companion simulation: http://rpubs.com/jjc/16567.
Joy Christian
Research Physicist
 
Posts: 2370
Joined: Wed Feb 05, 2014 4:49 am
Location: Oxford, United Kingdom

Re: The (still) open one-sided bet.

Postby gill1109 » Sun May 04, 2014 10:20 am

Joy Christian wrote:It is also important to not that the simulation I have posted is NOT Pearle's model.

Joy you didn't listen.

You may claim that a simulation of Pearle's model is also a simulation of your model, if you want to. I'm not sure who will believe you, but it is a logically allowable claim.

You cannot claim that the present simulation is *not* a simulation of Pearle's model, because it very definitely is, as anyone can see who is capable of reading both Pearle's paper and the present R code. Line for line, the code simulates Pearle's detection loophole LHV model.

It's irrelevant whether or not someone who is capable of reading both your works and the present R code can say that the code simulates your model.

Two pieces of computer code can simulate two different models if there is some mathematical isomorphism between the two mathematical structures. Wonderful! That's mathematics for you, showing that two things which a priori seem completely unrelated actually share the same underlying mathematical structures. That's precisely what mathematics is all about!
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby jreed » Sun May 04, 2014 3:01 pm

Joy Christian wrote:
In any case, I have updated the current simulation to reiterate the proper credits from the companion simulation: http://rpubs.com/jjc/16567.


It's difficult to keep up with these simulations. The one previous to this got rid of the "good" logic that looked like the detection loophole. Now it has returned to be used for "sets the topology of S^3". I programmed the previous version up in Mathematica, and I will also do that with this one.
jreed
 
Posts: 78
Joined: Mon Feb 17, 2014 5:10 pm

Re: The (still) open one-sided bet.

Postby gill1109 » Sun May 04, 2014 3:22 pm

jreed wrote:
Joy Christian wrote:
In any case, I have updated the current simulation to reiterate the proper credits from the companion simulation: http://rpubs.com/jjc/16567.


It's difficult to keep up with these simulations. The one previous to this got rid of the "good" logic that looked like the detection loophole. Now it has returned to be used for "sets the topology of S^3". I programmed the previous version up in Mathematica, and I will also do that with this one.

The topic is devoted to a challenge (it should not be called a bet, sorry, my mistake). Read the initial posts. The challenge is not about Christian's model or Pearle's model per se. Actually it is about (if it is about anything) Christian's colourful exploding balls *experiment*. At some stage of that experiment, two computer files are produced, containing (ordinary) directions in R^3.

The present topic, and the challenge, is to create two computer files, which do a particular job. The same job which Christian claims the files coming out of his experiment would do. The challenge is to submit (to me) two files which do that job. They can be just dreamed up out of your head, they can be by-products of a simulation of Pearle's, Christian's, Tom, Dick or Harry's model... It's a kind of "proof of concept" of Christian's *experiment*.

So please, everyone, keep on-topic: the still open challenge.

If anyone thinks they can win it by simulating Christian's model, they are welcome to go ahead. First come, first served. But please read the instructions first.
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby minkwe » Mon May 05, 2014 10:57 am

gill1109 wrote:The present topic, and the challenge, is to create two computer files, which do a particular job. The same job which Christian claims the files coming out of his experiment would do. The challenge is to submit (to me) two files which do that job. They can be just dreamed up out of your head, they can be by-products of a simulation of Pearle's, Christian's, Tom, Dick or Harry's model... It's a kind of "proof of concept" of Christian's *experiment*

I don't intent to push this on this thread, but it is interesting that I've challenged Richard with the flip-side of his challenge. He claims that QM/non-real/non-local theories can do something which LHV theories can not. He even claims that the definitive Bell experiment will be done within a few years. So I've asked him in the related thread to provide a "proof of concept" dataset from this definitive Bell-test experiment which violates the appropriate bound by just 0.00001. He is free to use non-locality/non-realism/statistical error to produce his dataset. He is free to just dream up the data. No restrictions about N, he is free to use any loophole he likes. We will use his Larsson & Gill theorem to determine the appropriate upper bound to compare with.
If anyone thinks QM can do something which LHV theories can not, they can help Richard to produce the non-local/non-real dataset.

The details are here: viewtopic.php?f=6&t=48&p=2021#p2017
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: The (still) open one-sided bet.

Postby gill1109 » Mon May 05, 2014 5:22 pm

minkwe wrote:
gill1109 wrote:The present topic, and the challenge, is to create two computer files, which do a particular job. The same job which Christian claims the files coming out of his experiment would do. The challenge is to submit (to me) two files which do that job. They can be just dreamed up out of your head, they can be by-products of a simulation of Pearle's, Christian's, Tom, Dick or Harry's model... It's a kind of "proof of concept" of Christian's *experiment*

I don't intent to push this on this thread, but it is interesting that I've challenged Richard with the flip-side of his challenge. He claims that QM/non-real/non-local theories can do something which LHV theories can not. He even claims that the definitive Bell experiment will be done within a few years. So I've asked him in the related thread to provide a "proof of concept" dataset from this definitive Bell-test experiment which violates the appropriate bound by just 0.00001. He is free to use non-locality/non-realism/statistical error to produce his dataset. He is free to just dream up the data. No restrictions about N, he is free to use any loophole he likes. We will use his Larsson & Gill theorem to determine the appropriate upper bound to compare with.
If anyone thinks QM can do something which LHV theories can not, they can help Richard to produce the non-local/non-real dataset.

The details are here: viewtopic.php?f=6&t=48&p=2021#p2017

How many Canadian dollars are you offering?
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby minkwe » Mon May 05, 2014 10:22 pm

gill1109 wrote:How many Canadian dollars are you offering?

I'm offering 1000000 Lunar dollars. And an opportunity for you to save-face and not have to withdraw halve a dozen papers claiming that QM/non-locality/non-reality can violate bounds which LHV can not. Now where is the dataset?
Last edited by minkwe on Mon May 05, 2014 10:57 pm, edited 1 time in total.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: The (still) open one-sided bet.

Postby gill1109 » Mon May 05, 2014 10:37 pm

Please get back on topic, Michel
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby minkwe » Mon May 05, 2014 10:57 pm

gill1109 wrote:Please get back on topic, Michel

It is your thread, you asked me a question, I answered it.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: The (still) open one-sided bet.

Postby FrediFizzx » Mon May 05, 2014 11:29 pm

minkwe wrote:
gill1109 wrote:How many Canadian dollars are you offering?

I'm offering 1000000 Lunar dollars. And an opportunity for you to save-face and not have to withdraw halve a dozen papers claiming that QM/non-locality/non-reality can violate bounds which LHV can not. Now where is the dataset?

You know he can't produce that dataset since it is impossible. Another thing that is most likely impossible; getting Richard to admit it publicly. But what the hey? Let him put out his wrong papers. It will come back on him eventually.
FrediFizzx
Independent Physics Researcher
 
Posts: 1978
Joined: Tue Mar 19, 2013 7:12 pm
Location: N. California, USA

Re: The (still) open one-sided bet.

Postby Heinera » Tue May 06, 2014 3:40 am

minkwe wrote:
gill1109 wrote:The present topic, and the challenge, is to create two computer files, which do a particular job. The same job which Christian claims the files coming out of his experiment would do. The challenge is to submit (to me) two files which do that job. They can be just dreamed up out of your head, they can be by-products of a simulation of Pearle's, Christian's, Tom, Dick or Harry's model... It's a kind of "proof of concept" of Christian's *experiment*

I don't intent to push this on this thread, but it is interesting that I've challenged Richard with the flip-side of his challenge. He claims that QM/non-real/non-local theories can do something which LHV theories can not. He even claims that the definitive Bell experiment will be done within a few years. So I've asked him in the related thread to provide a "proof of concept" dataset from this definitive Bell-test experiment which violates the appropriate bound by just 0.00001. He is free to use non-locality/non-realism/statistical error to produce his dataset. He is free to just dream up the data. No restrictions about N, he is free to use any loophole he likes. We will use his Larsson & Gill theorem to determine the appropriate upper bound to compare with.
If anyone thinks QM can do something which LHV theories can not, they can help Richard to produce the non-local/non-real dataset.

The details are here: viewtopic.php?f=6&t=48&p=2021#p2017


I don't understand what dataset you are talking about. What data should the set contain?

I can easily write a non-local hidden variables model that reproduces the QM correlations exactly. Would that be of interest to you?
Heinera
 
Posts: 767
Joined: Thu Feb 06, 2014 1:50 am

Re: The (still) open one-sided bet.

Postby minkwe » Tue May 06, 2014 4:44 am

Heinera wrote:
minkwe wrote:The details are here: viewtopic.php?f=6&t=48&p=2021#p2017

I don't understand what dataset you are talking about. What data should the set contain?

I can easily write a non-local hidden variables model that reproduces the QM correlations exactly. Would that be of interest to you?

My challenge is not about reproducing QM correlations. It is about violating Richards LG theorem by 0.00001. If you also believe non-local/non-real theories can easily do that. Then go ahead and produce the dataset which violates Richards LG theorem. The "proof of concept" for the loophole free Bell test experiment, you guys believe is around the corner. You guys believe QM is non-local/non-real and therefore the experiment will be non-local/non-real, don't you? You guys believe it is very easy for statistical error to violate bounds don't you? The experiment will produce data, won't it?

You can read further details in the link above.
minkwe
 
Posts: 1151
Joined: Sat Feb 08, 2014 10:22 am

Re: The (still) open one-sided bet.

Postby Heinera » Tue May 06, 2014 5:21 am

minkwe wrote: The experiment will produce data, won't it?

Yes, the experiment will produce a long 4xN table, where each line in the table looks like this:

AliceSetting, AliceObservation, BobSetting, BobObservation

Is this what you want me to simulate?
Heinera
 
Posts: 767
Joined: Thu Feb 06, 2014 1:50 am

Re: The (still) open one-sided bet.

Postby gill1109 » Tue May 06, 2014 5:39 am

Heinera wrote:
minkwe wrote: The experiment will produce data, won't it?

Yes, the experiment will produce a long 4xN table, where each line in the table looks like this:

AliceSetting, AliceObservation, BobSetting, BobObservation

Is this what you want me to simulate?


No we are talking about the coincidence loophole. And totally off-topic, here. Michel should open a new topic if he wants to discuss it.

The experiment produces a long 4XN table where each line looks like this:

AliceSetting, AliceObservation, AliceTime; BobSetting, BobObservation, BobTime

Each line is an iid observation from a very simple LHV model

Observations are only processed to give correlations if the difference between the two times is less than some threshhold.

http://arxiv.org/abs/quant-ph/0312035
Bell's inequality and the coincidence-time loophole
Jan-Ake Larsson, Richard Gill

This paper analyzes effects of time-dependence in the Bell inequality. A generalized inequality is derived for the case when coincidence and non-coincidence [and hence whether or not a pair contributes to the actual data] is controlled by timing that depends on the detector settings. Needless to say, this inequality is violated by quantum mechanics and could be violated by experimental data provided that the loss of measurement pairs through failure of coincidence is small enough, but the quantitative bound is more restrictive in this case than in the previously analyzed "efficiency loophole."

Europhysics Letters, vol 67, pp. 707-713 (2004)
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

Re: The (still) open one-sided bet.

Postby Heinera » Tue May 06, 2014 5:54 am

gill1109 wrote:
Heinera wrote:
minkwe wrote: The experiment will produce data, won't it?

Yes, the experiment will produce a long 4xN table, where each line in the table looks like this:

AliceSetting, AliceObservation, BobSetting, BobObservation

Is this what you want me to simulate?


No we are talking about the coincidence loophole. And totally off-topic, here. Michel should open a new topic if he wants to discuss it.

The experiment produces a long 4XN table where each line looks like this:

AliceSetting, AliceObservation, AliceTime; BobSetting, BobObservation, BobTime

Each line is an iid observation from a very simple LHV model

Observations are only processed to give correlations if the difference between the two times is less than some threshhold.

http://arxiv.org/abs/quant-ph/0312035

Yes, I've read the paper for sure, and I still don't understand what simulated data Michel is requiring to settle his bet (in the simulation I would just set AliceTime equal to BobTime. No coincidence-time loophole, and none needed, since the simulation I had in mind is non-local).

(OK, off topic; he can discuss his bet in a new thread).
Heinera
 
Posts: 767
Joined: Thu Feb 06, 2014 1:50 am

Re: The (still) open one-sided bet.

Postby gill1109 » Tue May 06, 2014 6:14 am

Heinera wrote:

Yes, I've read the paper for sure, and I still don't understand what simulated data Michel is requiring to settle his bet (in the simulation I would just set AliceTime equal to BobTime. No coincidence-time loophole, and none needed, since the simulation I had in mind is non-local).

You are right Heinera, it is very unclear what Michel wants. One can write down four lines of data such that the four empirical correlations equal +/- 1 (three negative and one positive, say). One can duplicate theses lines as often as one likes and add some lines with "non-coincidence" particle pairs, so as to get any gamma you like.

Because it was fun, I actually already wrote the code to simulate the model explicitly described at the end of our paper. It is completely local realistic and each particle pair is an independent sample from a fixed population of particle pairs. Each new pair of settings is selected randomly from the usual four pairs. None of the pairs are thrown away. The generated data are N lines of the form:

AliceSetting, AliceObservation, AliceTime; BobSetting, BobObservation, BobTime

One analyses the data just like people do in real experiments: keeping just those lines of the file, for which the two measurement times are within a certain fixed coincidence window. One calculates the empirical analogues of the four correlations and of the detection rate gamma. Then one looks at LG's inequality.

The model "saturates" (or attains) the inequality, ie in the large N limit one gets equality. With finite N, one has about 50-50 chance to exceed it. If N is not too large, the overshoot will often be much more than the tiny amount mentioned by Michel.

The code is ready, the data is ready, I am just wondering how much Michel will pay me for it, per line of data.
gill1109
Mathematical Statistician
 
Posts: 1943
Joined: Tue Feb 04, 2014 10:39 pm
Location: Leiden

PreviousNext

Return to Sci.Physics.Foundations

Who is online

Users browsing this forum: No registered users and 7 guests

cron
CodeCogs - An Open Source Scientific Library