|
@michael_nielsen | |||||
|
I am quite puzzled by this. I half wonder if I've simply messed up in relating Q and S, but I've checked the calculation twice, and I don't see it.
|
||||||
|
||||||
|
michael_nielsen
@michael_nielsen
|
3. velj |
|
A question for people who know about Bell inequalities: in the 1982 Aspect paper they present them in this form (a version of the CHSH inequality), with an experimental value (& violation) of S = 0.101 pic.twitter.com/MR1CcBfO2q
|
||
|
|
||
|
michael_nielsen
@michael_nielsen
|
3. velj |
|
The more commonly stated form of the CHSH inequality is Q := E(AC)+ E(BC) + E(BD) - E(AD) <= 2, where A, B etc are the observables for the polarization (values +- 1). By my calculation, Q = 2+4S, so the experimental value in Aspect corresponds to Q ~ 2.404.
|
||
|
|
||
|
michael_nielsen
@michael_nielsen
|
3. velj |
|
This is well short of the maximal violation (Q = 2 sqrt(2) ~ 2.8). But AFAICT the Aspect experiment uses polarization angles that should give the maximal violation
What gives? Have I goofed? Or am I missing something? Is it the cryptic remark about solid angles & efficiencies? pic.twitter.com/w4uhbfiEq5
|
||
|
|
||
|
michael_nielsen
@michael_nielsen
|
3. velj |
|
Original paper: journals.aps.org/prl/pdf/10.110…
|
||
|
|
||
|
Matt Leifer
@mattleifer
|
3. velj |
|
It can't be detector inefficiencies because they make the fair sampling assumption. My guess is dark counts plus events where they only detect one of the two particles.
|
||
|
|
||
|
michael_nielsen
@michael_nielsen
|
4. velj |
|
Thanks Matt.
|
||
|
|
||
|
Andrew M. Webb
@AndrewM_Webb
|
3. velj |
|
I'm not an expert, but I was able to find this page discussing the paper which says (is Sec. 7.1.3) that while the theoretical maximum is ~2.8, correcting for "instrumental deficiencies" the maximum is 2.448. I can't see where they've got that number from. ugr.es/~bosca/WebFCen…
|
||
|
|
||
|
Andrew M. Webb
@AndrewM_Webb
|
3. velj |
|
Well, I guess the number agrees with the 0.112 maximum of S given in the paper, but I can't see any reference to it being a correction for instrumental deficiencies
|
||
|
|
||
|
Andrew Doherty
@andrewcdoherty
|
3. velj |
|
My recollection is that due to detector inefficiencies and source imperfections these experiments were indeed quite far from the maximal violation
|
||
|
|
||
|
michael_nielsen
@michael_nielsen
|
4. velj |
|
Thanks Andrew. Matt makes the following argument, which seems likely right, about inefficiencies: twitter.com/mattleifer/sta…
|
||
|
|
||
|
Danielle Fong🦠🧬💨🇹🇷🛫🐉💗🕵🏻♀️
@DanielleFong
|
3. velj |
|
Giving cuckoo’s egg vibes.
|
||
|
|
||