The h Bar

enumaris

16:03

wahhhh

Anonymous

@Semiclassical BTW did you complete writing that paper on the correlation and Bell's inequalities stuff?

Anonymous

Oh, found the link: overleaf.com/read/njgmgntgjvrj

Semiclassical

not yet, no

sorta lost track of it

(beyond those notes, there's eventually a paper that'll come out. but the main prof working on it has been busier than he expected, so it's fallen a bit by the wayside)

Anonymous

I can understand. Do you have any previous papers in this quantum information area though?

Semiclassical

nooope

this was my first thing

Anonymous

16:09

Ah, cool cool :)

Semiclassical

this was just a fun side project, not research with my advisor

Anonymous

@Semiclassical I see! And are you still looking for industry jobs or planning on staying in academia?

Semiclassical

i'm pretty stuck in that regard

Anonymous

Uh, why? :P

Semiclassical

I've been a student so long that it's been tough for me to break out of that mindset

I've basically been in one academic program or for the last 25 years (elementary school - grad school)

Anonymous

16:13

"one academic program"?

Semiclassical

"one academic program or another", woops

glS

before I go on and ask a stupid question on the main site, let me try here. Consider a time-dependent Hamiltonian $H(t)$, and let $|n\rangle$ and $|m\rangle$ be two (time-dependent) eigenstates of $H$ with the same energy $E$. Then, differentiating the eigenvalue equation $H|n\rangle=E|n\rangle$ and applying $\langle m|$ to the result, we get $\langle m|\partial_t H|n\rangle=0$, which is clearly not true. Where is the catch?

Semiclassical

i.e. K12, then college, then grad school

Anonymous

@Semiclassical That's fine I guess. Lots of people complete their PhDs at 33 or more! Never late to try new things :)

Semiclassical

@glS since they're time-dependent states, shouldn't you have $\partial_t (H|n\rangle) = (\partial_t H)|n\rangle +H(\partial_t |n\rangle)$?

glS

16:15

@Semiclassical indeed, but if you do the math, and both states have the same eigenvalue, the $\partial_t |n\rangle$ terms cancel out

Semiclassical

hmm

glS

this is the general formula, which is true in the nondegenerate case (and is what is used for example to derive the adiabatic theorem)

Semiclassical

well, note that you'd need $\langle m | n\rangle =\delta_{nm}$

glS

indeed, but that should be true, because you can choose orthogonal degenerate eigenstates, right?

Semiclassical

You typically can when the system is time-independent

I'm not at all sure you can do so when it's not.

glS

16:19

@Semiclassical but here $|n\rangle$ and $|m\rangle$ should be the instantaneous eigenstates of the system, so surely at any given time we can pick two orthogonal ones?

Here is a trivial example of this in action: consider the Hamiltonian $H(t)=t \sigma_x$. Then $|0\rangle$ and $|1\rangle$ are degenerate for $t=0$ but the degeneracy lifts for $t>0$. The eigenstates are time-independent in this case. You have $H'(t)=\sigma_x$, so if you differentiate $H|0\rangle=t|0\rangle$ you get $|1\rangle=|0\rangle$, which makes no sense?

Semiclassical

But $|0\rangle,|1\rangle$ aren't time-dependent eigenstates of $H(t)$

glS

ah, I think I got it. The problem is in this example that the eigenstates are time-independent and equal to $|\pm\rangle$, not $|0\rangle,|1\rangle$. So the problem is that if we make the wrong choice of inititial nondegenerate eigenstates, you get that $|E_i(t)\rangle$ are non-differentiable at the moment the degeneracy lifts... does this make sense?

Semiclassical

they're time-dependent eigenstates of $t\sigma_z$

well, let's see. we can get rid of that issue by doing $H(t)=t\sigma_z$

oh, you have $E_n=E_m$ only at $t=0$ in that case. hmmmm

well, at $t=0$, the $|\pm\rangle$ states are still eigenvectors

I think I agree, though.

The point I was going to make earlier was the following. Suppose you've got time-dependent states $|n(t)\rangle, |m(t)\rangle$ which are degenerate eigenstates of $H$

the trouble is that, in order to preserve orthogonality, $A$ and $B$ will be functions of the time $t=\tau$

enumaris

whoop, answered a GR question for the first time in a while

I feel accomplished

Semiclassical

as such, there's no guarantee that $A(t)|n(t)\rangle+B(t)|m(t)\rangle$ will itself be an eigenstate. the time-derivatives will act on the coefficients

glS

16:33

@Semiclassical I'm not sure I follow. Here $|n(t)\rangle$ is the instantaneous eigenstate of $H(t)$?

Semiclassical

Right.

the bottom line is that, while you can certainly choose your degenerate states to be orthogonal at a given instant, you can't demand that they remain orthogonal for all time

the condition for orthogonality can vary with time

glS

@Semiclassical so you are following the eigenenergies from a time in which they are nondegenerate back to the point in which they are degenerate, and looking at how the eigenstates evolve?

Semiclassical

I'm assuming that they remain degenerate for all time in this calculation

I'm not entirely sure I buy what I'm saying, though.

An example would be nice.

glS

@Semiclassical ok, but then why can't you just define them as always orthogonal? After all, if they stay degenerate, at any point in time you can find a pair of orthogonal states in the degenerate eigenspace

@Semiclassical an example of two states remaining degenerate? What about simply $H=0$?

Semiclassical

eh, that seems too trivial to be much help here.

Note that, if you took your basis in the above case to just be $|\pm \rangle$, then you have $t\sigma_x |\pm \rangle = \pm t |\pm \rangle$

in which case differentiation gives $\sigma_x|\pm \rangle = \pm |\pm\rangle$ which is true

glS

16:43

@Semiclassical yes that is what I noticed above, when saying that the problem seems to be in a "wrong" choice of initial states in the degenerate eigenspace

Semiclassical

The tricky thing with this example is that there's a basis in which the states aren't time-dependent

glS

ah, so maybe it is true that $\langle n|\partial_t H|m\rangle=0$, provided we choose $|n\rangle$ and $|m\rangle$ to be those states which continuously evolve into the nondegenerate eigenstates

Semiclassical

Yeah, I think that's right

A good example to play around with is probably something like $H=H_0+t H_1$

which is a common enough example in such problems anyways

(I know there's a name for that kind of problem but durned if I can remember it)

right, Landau-Zener

glS

problems studied in perturbation theory?

Semiclassical

Landau–Zener formula

The Landau–Zener formula is an analytic solution to the equations of motion governing the transition dynamics of a 2-level quantum mechanical system, with a time-dependent Hamiltonian varying such that the energy separation of the two states is a linear function of time. The formula, giving the probability of a diabatic (not adiabatic) transition between the two energy states, was published separately by Lev Landau, Clarence Zener, Ernst Stueckelberg, and Ettore Majorana, in 1932. If the system starts, in the infinite past, in the lower energy eigenstate, we wish to calculate the probability...

glS

16:46

@Semiclassical ah, yes that is the case $H=\sigma_z+\Gamma(t)\sigma_x$ if I remember correctly.

Semiclassical

yeah, that's the classic Landau-Zener case

my advisor was having me look at some solvable multistate L-Z problems at one point, but I didn't get far

glS

yea ok that is a good example. The gist seems to be the nondifferentiability of the eigenstates at the degenerate point, if they are not chosen carefully. That is actually an interesting aspect

Semiclassical

yeah

the trouble is finding examples which are nontrivial but not impossible tho

and the fact that the scattering amplitude of the Landau-Zener problem has an exact solution is a strange magic

(the fact that certain multistate problems are solvable is even weirder.)

more generally there's some connections with so-called integrability of time-dependent Hamiltonians: arxiv.org/abs/1711.09945

glS

@Semiclassical that looks very interesting, thanks

Semiclassical

hmm, something I'm trying to formulate

We label the result of the L-Z problem as a scattering amplitude

Is there a way to make that more-or-less literal, i.e. is there a way to have that scattering amplitude show up in an actual scattering process?

glS

17:03

@Semiclassical are you referring to the paper? I'm not sure what you mean by "labeling results as scattering amplitudes"

Semiclassical

I think I'm flashing back to the title of this paper when I say that: arxiv.org/abs/1609.06285

with the scattering in that sense being the case that the incoming and outgoing states will be asymptotic to states of $H_1$, with there being some definite transition probabilities between the various options

But in a scattering process one is talking about initial/final states in space, not in time

which is what has always weirded me out with that language

glS

@Semiclassical I'm lost as to what is the question here =)

Semiclassical

lol

Suppose you've got a Hamiltonian $H=H_0+tH_1$

glS

@Semiclassical isn't that the same? they are assuming that for $t=\pm\infty$ the states are not interacting?

Semiclassical

sure, but the Landau-Zener formula is formulated purely in terms of time

there's no space at hand

I guess what I'm trying to think about is how one maps an adiabatic problem to a literal scattering problem

glS

17:20

@Semiclassical well in a scattering problem you have kind of the opposite situation right? a sudden change at some point. Do you have degeneracies in these kinds of situations?

Semiclassical

that's a good way to think about it

I guess what comes to mind from that POV is if you were scattering off a spin-dependent potential step

e.g. if spin-up then $V(x)=0$ if $x<0$ and $V(x)=V_0>0$ for $x>0$, and vice-versa for spin-down

to make the connection with Landau-Zener one probably needs to replace a sharp change at $x=0$ with a gradual change

Yellow

17:34

does anyone have irodovs problems in general physics typesetted in latex?

Cows

:)

vzn

@Blue cool did semiclassical write that? are you interested in Bells thm?

Semiclassical

Yeah, it's mine

Anonymous

@Yellow There are newer reprints of Irodov but I don't know if you'll get the LaTeX version.

Semiclassical

Have not updated it in weeks (months?)

vzn

17:41

@Semiclassical something to look into this area (maybe revisit), CHSH analysis

Semiclassical

yeah, that's in there as well

or at least a touch of it is

can't remember how far I got in those notes

Yellow

@Blue on libgen all of them are from the same publisher MIr

vzn

@Semiclassical CHSH is (very) rarely reexamined.

glS

@Semiclassical I mean I don't really see the connection with the previous case though. If you also introduce discontinuity in the Hamiltonian/potential then you have a second-order discontinuity of the eigenstates, but apart from that the rest still holds

@vzn what's wrong with CHSH analysis?

Yellow

have you guys read landaus mechanics book

vzn

17:43

@glS didnt say there was something wrong. its very subtle. has some ("hidden") "assumptions".

Semiclassical

basically what I'd like to do is map an adiabatic transition problem to a scattering problem

Yellow

under the courses of theoretical physics

Anonymous

@Yellow Buy this one. It's quite cheap

glS

@vzn such as?

Semiclassical

which I can't imagine is that hard. my brain just isn't working

Yellow

17:43

@Blue pretty sure I will get it in the olympiad camp

vzn

@glS its close to/ centered on the "fair sampling" issue.

Yellow

just wanted to do some problems without my eyes bleeding

have to wait 3 days

Semiclassical

Subject to the assumptions they make, Bell's thm and its descendants are sound.

glS

@vzn what do you mean by "centered" on the fair sampling issue?

Semiclassical

But that premise is important.

vzn

17:44

@Semiclassical the assumptions are (maybe) bigger than recognized/ admitted.

Anonymous

@Yellow There's a website which contains all the Irodov problems

Yellow

nice

Anonymous

I forgot the URL but you'll get it if you look for it

Semiclassical

eh. i'm not sure that's true, given how much talk there's been about the various loopholes

Anonymous

exir.ru/solutions

vzn

17:45

@Semiclassical theres a lot of talk about so called loophole free tests also, and dont (entirely) agree with that designation.

Anonymous

Solutions are in Russian

Semiclassical

I don't know enough about them to say anything, so I won't :P

vzn

@glS its listed here, but maybe deserves more analysis/ attn than has gotten in past. ie revisiting/ reexamination. en.wikipedia.org/wiki/Loopholes_in_Bell_test_experiments

glS

I mean I don't know, I personally see CHSH and company as statements about certain conditional probability distribution not being possible under some assumptions on the correlations between some of the variables

Semiclassical

@glS same

Yellow

17:46

google translate works for words

glS

@vzn you are still being very vague here

vzn

@glS as does nearly everyone else.

Anonymous

@Yellow Yeah, but not for images

Yellow

amazon.com/Mechanics-Course-Theoretical-Physics-Landau/dp/…

vzn

@glS am pointing to a very specific paper and specific issue. ppl around here seem to call "vague" anything that isnt fully laid out. already said its subtle

glS

17:47

@vzn the loopholes are problems in the physical interpretation of the inequalities, not on the inequalities themselves

vzn

@glS dont fully agree with that. but must say youre one of the few to not dismiss it immediately.

Semiclassical

@glS or at least on the physical interpretation of the experiments that've been done

fair sampling being a big one

glS

@vzn but if you say a problem is "subtle" without saying exactly why, and then giving a quite generic reference to support the statement, that makes it look like you yourself don't know exactly what you are talking about.The loopholes are again a well-known aspect of Bell's inequalities and many people work on those, what is subtle about it?

vzn

think there is a potentially large issue that has been somewhat lost or buried in the sands of time. it would require more than short-term attention (ie patience) to further isolate. am not inclined to respond/ involve myself with those who dont have any/ jump to (negative) conclusions etc.

Semiclassical

Cuius rei demonstrationem mirabilem sane detexi hanc marginis exiguitas non caperet.

3

Anonymous

17:52

And here we are: back to square one (as usual) :P

glS

@Semiclassical ah, I knew I should have studied latin properly in high school!

vzn

@Semiclassical lol a very subtle area. FLT was proven after 3½ centuries and wiles just won abel prize for it.

Anonymous

Sure, but what's the "claim" you want to make about CHSH? Fermat at least had a statement :)

Anonymous

Stating that it's subtle is not much of a claim I'm afraid

glS

@vzn look, I'm not saying there is nothing to be done or discovered in these matters. To be honest I myself have never been able to fully "accept", on an intuitive level, the implications of Bell's inequalities and such, and I find myself often trying to make sense of it. But this doesn't mean that well established facts about Bell's inequalities are just "wrong", or that you were not being very generic in your claims

Anonymous

17:56

The interpretations of Bell's inequalities are interesting to ponder about, yes

Semiclassical

I mean, there's various ways in which Bell's inequalities could be "wrong"

one is that the theorem does not follow from the assumptions it makes. but I don't think there's really anything to challenge there; a proof is a proof.

another way is to dispute whether the assumptions it makes are sound. one can either do so on theoretical grounds, e.g. some rejection of counter-factual definiteness

or one can do so on experimental grounds, i.e. that the experiment you're doing doesn't actually fulfill the assumptions being made

Yellow

Does it makes sense to talk about a metric on a circle?

Anonymous

6

Q: Tracking down the locality assumption in a CHSH inequality derivation

CHSH inequality requires both locality and realism. I will equate here realism with counterfactual definiteness. Now counterfactual definiteness tells us that given two different measurements on the same object, described by random variables $C_1$ and $C_2$, there exists a joint probability dist...

Anonymous

There was a thread on this

Anonymous

And a few more

glS

17:59

@Blue I don't like Ron's answer there. I prefer the answer to the linked question =)

Semiclassical

Along the lines of the last one, you can challenge the extent to which a given experiment addresses the known loopholes

e.g. you might fairly claim that a given experiment validates Bell's inequality assuming detector efficiency of at least X

I think such claims are particularly important when it comes to various modifications of QM, in particular ones where it's claimed that QM is only valid as a limiting case of another theory

I have in mind stuff like GRW and other objective collapse theories

Anonymous

@glS Ah, nice! I hadn't noticed that it was written by you

glS

@Semiclassical although I really don't think the loopholes is where one should look for "problems" in Bell's theorem. I mean if you have to suppose an extremely magical and unlikely kind of dynamics that somehow just happens to reproduce the results of QM with LHVs, then you might as well just assume the superdeterminism (or whatever it's called) hypothesis and be done with it

Anonymous

I was actually looking for some resource recommendations for the Bell inequality derivations last month. I didn't find much

Anonymous

Wikipedia is good, but I don't yet understand all the assumptions used

Semiclassical

18:05

@glS eh, I don't think the models are really all that magical or strange. for instance, the de Broglie - Bohm account is remarkably simple and internally consistent, so long as one doesn't worry about relativity :P

vzn

@glS the claims are not generic and invite anyone serious to inquire deeper but its very timeconsuming, have spent decades on this and the end is not yet in sight. have little patience for the personal/ nearly ad hominem noise in these chat rooms (alas, somewhat typical/ routine at times) chat.stackexchange.com/rooms/9446/theory-salon ... have been careful not to claim any "conventional wisdom" is wrong

Semiclassical

my own disquiet is more to do with how (un)productive various interpretations are

glS

actually, given that we are on the subject, as anyone ever heard of connections made between Bell's inequalities and the Fisher information? I ask because it seems to me that nonlocal but still signalling theories (such as QM) are exactly those corresponding to vanishing Fisher information between Alice's measurement choice and Bob's result

@Semiclassical Is pilot wave considered to be a "loophole"? I just thought of it as an alternative interpretation of QM

Semiclassical

I'm fond of the deBB account, for instance, but it's hard to deny that it's been remarkably irrelevant to the progress of quantum theory

@glS you're right, of course---it's not an LHV, so it can (and does) work as an alternative account

what I have in mind is more the various modifications that the pilot-wave account has inspired, e.g. the sort of pilot-wave hydrodynamics that Bush-Couder want

vzn

@Semiclassical true on one level on other hand try looking into the deep history/ story of von Neumanns impossibility proof. it seems to be rarely noted/ taught/ mentioned these days. ("sands of time")

Semiclassical

18:09

I don't think those models are especially contrived or 'magical.'

@vzn i've been meaning to reexamine that myself. the conventional wisdom at this point: von Neumann gave a proof of hidden variables being impossible, which everyone accepted as right b/c hey von Neumann is smart and HVs are infantile. But then Bell comes along and points out that the assumptions von Neumann made are silly.

glS

@Semiclassical I myself don't know much about it, never really found the time to properly understand it. But I have to say that I still think it is just as magic as the regular interpretation of QM, in that it shares (as it of course has to) the property of allowing nonlocal correlations while still be no-signalling. I really cannot wrap my head around how nonlocal, non-signalling theories make physical sense

vzn

@Semiclassical not disagreeing with anything you said, thx for nice summary; however the assumptions were not so "silly" at the time he made them. they were only "silly" with 2020 hindsight that eluded many of the top minds of the era. ie... subtlety

Semiclassical

@vzn well, I chiefly have in mind Bell's remark that von Neumann's theorem was 'not only false but foolish'

More recently, there's an article by Jeffrey Bub that challenges the history that Bell put forward

vzn

@Semiclassical bell found the glitch and was deservedly triumphant about it (but he did tend to have a nearly "trash talking" side.) anyway it was a mini-kuhnian shift.

Semiclassical

the point namely being that, whatever anyone else may have done, von Neumann never claimed to have proven more than he did

(my own sense is that, while Bell might have not been doing justice to von Neumann's own thinking, he was probably responding to how other people were using von Neumann's proof. so the 'not just false but foolish' is really more a commentary on how von Neumann's impossibility theorem was received by the mainstream community. iirc the Bub paper really doesn't address this)

Tanuj

18:14

@Semiclassical Hey man , I've got something to ask.

glS

@Blue for which case you mean? To better understand Bell's inequalities I would recommend the same paper I mentioned in the question, the Brunner et al. review

vzn

@Semiclassical maybe ran across that in here after cited, maybe by you. need to look further at that. believe the reexamination is not yet finished. reexamination is not something to be kneejerk dismissed. to me thats some of the lesson of the von Neumann-Bell episode.

Semiclassical

@vzn it's this paper: arxiv.org/abs/1006.0499

vzn

@Semiclassical it is not always so obvious what proofs actually/ precisely prove, that is another lesson of the episode. there can be gaps in translating the formalism into words. with the bell analysis, that tricky word is nonlocality.

Semiclassical

In particular I myself would like to know what von Neumann's result actually has to say about deBB

something to do with contextuality or nonlocality, probably

vzn

18:18

@Semiclassical (yes familiar) searching, you cited it aug 2, think read abstract then, ok have now bookmarked & will look at some further.

Semiclassical

I haven't looked at that paper in quite a while, I'll note (though at the time I do remember it coming up in discussion at the physics-philosophy interest group)

Anonymous

@glS Umm, like for the CHSH they use the realism, locality and freedom of choice of measurements assumptions but it was not clear to me what happens when one of them is dropped

Anonymous

That looks like a really good paper btw

Semiclassical

So my recollection of that paper shouldn't be read as either an endorsement or rejection of its conclusions :P

Anonymous

I will read it when I get time

Semiclassical

18:21

@glS yeah. I mean, it depends on quite what we're talking about. if one is only worrying about deBB in the context of the Schrodinger equation, then I really do like the story you get from it.

vzn

anyway (coincidentally my thoughts return to CHSH also) my simple observation is that the CHSH "proofs" have even more moving parts than even von Neumanns proof and think (somewhat similarly/ analogously) something subtle yet lurks there, not yet pinpointed. in a way, nobody is really looking very hard there.

Semiclassical

the trouble is that...well, that's an emphatically non-relativistic story.

so that story being nonlocal isn't really a problem. the real problem is that the issue persists when you try to formulate a relativistic deBB account

vzn

@Blue those concepts are translated into mathematical properties and again, sometimes there can be subtle gaps between concepts and formalism, and again cite von Neumann proof as an example. people eg Bub are still arguing about that exact link/ connection between the words and formalism over ½ century later. thats kind of staggering.

glS

@Blue yes. I myself find it much clearer to just say that it assumes that the outcomes are described by conditional probability distributions of the form $p(ab|xy)=\sum_\lambda p(a|x\lambda)p(b|y\lambda)p(\lambda)$, which summarises all of the assumptions: 1) realism is the existence of the "hidden variable" $\lambda$, the knowledge of which would allow to describe the outcomes as uncorrelated, 2) locality is the assumption that $p(ab|xy\lambda)=p(a|x\lambda)p(b|y\lambda)$

Semiclassical

at a technical level, it still works out. but there it really does start to feel ~~magical~~ not the word I mean

glS

18:24

and "freedom of choice" is the assumption that $p(xy)=p(x)p(y)$, that is, measurement choices are not correlated

Semiclassical

I guess you could say: It starts to feel like you're forcing the theory into a picture that it doesn't want to be fit into

Anonymous

@glS Interesting! I was looking exactly for something like this which explains the necessity of each and every assumption clearly when translated to the mathematical formulation

Anonymous

I'm a bit out of touch with the derivations, but this is surely something I'd like to acquaint myself at a greater depth later on

Semiclassical

Though for me the real embarrassment of deBB is not its nonlocality or even its issues with relativity

it's how utterly irrelevant it is for actual physics problems

Anonymous

@glS Ah, basically freedom of choice indicates independence of the distributions

Anonymous

18:27

Makes sense

Semiclassical

sure, you can take a concrete experimental problem and show how Bohmian trajectories yield the QM results

vzn

btw re "sands of time," dont recall when just found this, ran across this recently, maybe on dialog or link in here, not sure. apparently Grete Hermann played a key role in spotting that von Neumann proof loophole, maybe Bell was unaware? need to learn more on this. philsci-archive.pitt.edu/12443

Semiclassical

But at an experimental level, how much have you actually learned/

@vzn I think the usual history is that she noticed the loophole but her paper didn't catch people's attention

vzn

@Semiclassical have read many histories over the years and she is not even cited in very many, only found out about her recently. but the von Neumann proof is not even mentioned in a lot of histories either.

glS

@Semiclassical but even in the non relativistic case, you still have a theory that describes a magical form of nonlocal correlations that somehow are just so perfectly tuned to not allow for nonlocal signalling. My problem with QM, Bohm or otherwise, is really just this "coincidence". I mean if you tell me that a theory is nonlocal I would be fine with that, it would just mean that we have to work out the reasons why it's consistent with relativity and all that. But QM is weirder than that

Semiclassical

18:30

in most of the stuff I've seen she is mentioned, but only that

vzn

@Semiclassical & btw its obviously "not her fault" if her paper wasnt noticed... its up to/ "on" everyone else to get their citations straight... it even has a little feel to the way that bohm was intentionally ignored by the larger community.

Semiclassical

didn't say it was

@glS to put it a little differently: local interactions are fine. visible nonlocal interactions are uncomfortable but at least thinkable. but hidden nonlocal interactions seem magical.

glS

@Blue yes. To be fair, I should mention that some would not agree with my interpretation of this, because it is not clear that you can talk of the "probability distribution of the experimenter making a given choice". But I much prefer this to saying very vaguely that "experimenters have free will and thus their measurement choices are independent"

ACuriousMind

@glS I don't think that is surprising: Since we know any more fundamental theory must have classical relativistic mechanics that is free of such signaling as its limit, it is hard to see how any theory with non-local signaling could ever describe our world.

vzn

@Semiclassical hidden nonlocal interactions are now proven to exist in classical mechanics by multiple authorities (eg La Cour-Ott) for years, and its our intutions that are off and need to be revised. but these findings do not fit with the status quo aka dominant paradigm and are not taken seriously or are dismissed right now.

Semiclassical

18:35

At a technical level, the way that deBB as I know of it achieves consistency with relativity is to require a 'dynamically privileged foliation of spacetime'

glS

@ACuriousMind sure, relativity is the reason why we don't like no-signalling. What I'm saying is that it seems like a pretty big "coincidence" that QM manages to be no-signalling while still being nonlocal (to be fair this is not a specific property of QM; there are other possible theories that are nonlocal and no-signalling)

Semiclassical

which philosophically goes against the whole idea of "relativity of simultaneity"

ACuriousMind

@glS I don't understand where the "coincidence" is. All the possible theories that have non-local signalling simply can't describe our world.

That a theory we use to describe our world is able to describe our world is not a coincidence!

Semiclassical

if one is doing non-relativistic QM, then the fact that simultaneity must be absolute is not really an issue to me. but in the deBB account one really seems to have to reject that

Sir Cumference

Mother of god it's 101°F here (38.3°C)

Semiclassical

18:37

which is hard for me to swallow

Sir Cumference

And this is way above the equator

Semiclassical

oof

glS

@ACuriousMind let's put it this way. I haven't been able to wrap my head around the possibility of correlations that are nonlocal while still no-signalling, which is the kind of thing QM produces

Semiclassical

@glS I think the way I'd formulate this point on the deBB side: It's not so hard to achieve technical compatibility of deBB with relativity. But it seems far harder to achieve this in a 'principled' way.

ACuriousMind

@glS That's an entirely different statement! I think the reason many find it hard to accept is because it seems to defy "causation" (in the folk sense of cause and effect) while still maintaining locality/no-signalling. We need to face that causation has never been a coherent concept to begin with :P

Semiclassical

18:41

@ACuriousMind How very Humean of you :P

vzn

emergent properties are hard to wrap ones brain around. but there are many in classical physics... the whole area was not studied seriously/ as a whole until over ½ century after invention of QM, one milestone is the Lorenz eqns, which were regarded as nearly revolutionary at the time

glS

@Semiclassical I have no idea what that means! Does deBB need to bring relativity in the argument even in this simpler case?

Semiclassical

@glS nah, I wasn't saying that

or, well

hmm

ACuriousMind

@Semiclassical Hume was a wise man

Semiclassical

This is a point where i"m actually not so sure on

Bernardo Meurer

18:43

@ACuriousMind o/

Semiclassical

I mean, at some level, one shouldn't expect to get true non-signalling in non-relativistic QM, if only because $c$ doesn't show up in the theory

ACuriousMind

@BernardoMeurer \o

Bernardo Meurer

Long time no see :)

What's up?

How's everything at SAP

ACuriousMind

@BernardoMeurer It's going fine. Work's good, and life's good, too. Although I did break my ankle :P

How about you?

Bernardo Meurer

Jesus, how?!

Semiclassical

18:45

for instance, if you prepare a particle in a sufficiently narrow state, then (according to non-relativistic QM) you'd have no reason to expect $\sigma_p<mc$. that's just the uncertainty principle in action

ACuriousMind

@BernardoMeurer I just tripped and twisted it in a very unfortunate way, apparently.

Bernardo Meurer

Good lord

That sucks

Is it better?

ACuriousMind

Yeah. 6 weeks on crutches :/

Bernardo Meurer

Damn :/

ACuriousMind

It doesn't hurt anymore but I'm not allowed to put my weight on it yet, so crutches it is

vzn

18:46

@BernardoMeurer hey dude missed you what have you been doing for months? having a life™? :P

Bernardo Meurer

@vzn Hey! Just working a lot :)

Worked all summer at an AI startup

Semiclassical

the reason we don't worry about that in nonrelativistic QM is because we know that, for instance: if you were to localize an electron to that extent, you'd start creating particles-antiparticles

ACuriousMind

At least my arms are getting a workout :P

Bernardo Meurer

Also @ACuriousMind they want me to hire you :P

vzn

@BernardoMeurer kickass! what are they doing? lots of ML talk around here while you were gone by enumaris + daneilunderwood

ACuriousMind

18:47

@BernardoMeurer An...AI startup? I think I could do without being dissected by amateurs :D

Bernardo Meurer

@vzn standard.ai

Automated checkout

Semiclassical

@gls I guess my upshot being that I'm not sure how one implements non-signalling (in the sense of not allowing observable velocities $v>c$) in non-relativistic QM without appealing to relativistic QM in some sense

Bernardo Meurer

@ACuriousMind Lol

vzn

@BernardoMeurer wow! thats a big deal, amazon has it going, are they affiliated at all?

Bernardo Meurer

Seriously though, we've been looking for a good SWE that could work with me and Brandon (one of the founders) on our low-level stuff (all written in Rust, almost no C)

@vzn Nope, we're their direct competitor

vzn

18:49

@BernardoMeurer lol ok good luck with that! o_O

Semiclassical

you need some sort of relativistic QM to quantify in what regime you expect non-relativistic QM to be an appropriate description

danielunderwood

I tried Rust once...it was interesting

Bernardo Meurer

@vzn We're already winning

Semiclassical

and I think the same applies to deBB as well. it's just more front-and-center

vzn

@BernardoMeurer am sure with you on the team :P what kind of ML + libraries etc do you work with... is it a lot of image/ video analysis? humans picking stuff off of shelves?

glS

18:50

@ACuriousMind I'm not so sure.. I totally agree and don't find particularly surprising that "causation" is a human conception. It's just our way to say "A is correlated with B and thus manipulating A I can 'cause' B". However, this kind of statement is purely about correlations, not causations. You have two things that can be correlated in a completely unexploitable way, which I find weird

Bernardo Meurer

@danielunderwood Rust has slowly turned into my favorite language (was C99 before)

ACuriousMind

@BernardoMeurer I've heard good things about Rust but I've yet to try it.

Semiclassical

@glS i tend to agree with this.

ACuriousMind

There's a Rust meetup by some of my colleagues that has a knack for always happening at times I can't possibly make it :P

Bernardo Meurer

@vzn For ML we use Tensorflow, like everyone else. I hand-wrote all our video-managing stuff in Rust. Some other things like cropping, etc are done in python by the ML devs

danielunderwood

18:52

I will say that I really liked the build/dependency/docs/test system (cargo I think?)

Bernardo Meurer

@danielunderwood Oh, yeah, Cargo is a bliss

danielunderwood

Especially compared to C/C++, but I think I preferred cargo even to python

Bernardo Meurer

SAT-based dependency managers <3

vzn

@BernardoMeurer talked to DS lately? havent seen him in ages either figured you guys were out partying on the beach or cleaning qubits or something :P

Bernardo Meurer

Hehehe, I haven't seen him in a while since I was living in SF over the summer; but we'll be having dinner today or tomorrow I think

vzn

18:53

@BernardoMeurer ? SATisfiability problem?

Semiclassical

I mean, one response is similar to Bohr: what's essentially quantum is the indivisibility of quantum phenomena from their measurement conttext

glS

@Semiclassical that is an interesting point. I'm tempted to say that no-signalling should mean that A's outcomes cannot be correlated with B's, but that is exactly the case we don't have in QM

Semiclassical

“the impossibility of any sharp distinction between the behavior of atomic objects and the interaction with the measuring instruments which serve to define the conditions under which the phenomena appear."

Bernardo Meurer

@vzn Boolean satisfiability, ye

vzn

@BernardoMeurer tell him hi for me & that there is copious feedback on his last paper awaiting when he returns :P

Bernardo Meurer

18:55

Hahaha, I will!

vzn

@BernardoMeurer is it part of your analysis/ product?

Semiclassical

One way to take the deBB story is: If you take this indivisibility from the get-go, then any attempt to formulate a spacetime story must take into account the entire measurement context from the start

Bernardo Meurer

Not yet, but the dependency/package manager for Rust (Cargo) uses SAT to solve for which dependencies to fetch

Semiclassical

and on the whole it demands one take the notion of an 'instantaneous configuration' quite literally

vzn

@glS one might say that the correlations are indeed being exploited... with QC + other applications etc!

Semiclassical

18:56

once you do that, you do get a story and one that's internally/experimentally consistent

but that seems like a rather big fish to swallow

(I can't remember the right metaphor)

glS

@vzn well no, that's an entirely different story. For QC you are not exploiting nonlocal correlations but rather the more complex dynamics allowed by the larger space in which the dynamics happens. Nonlocality is exploited for communication, QKD etc. But then again this is a different use of the word "exploit" than the one I was using before

Semiclassical

for better or worse, the deBB story is definitively a story about 3N-dimensional configuration space

vzn

@glS disagree, think nonlocality is a key property exploited by QC, but it isnt presented that way in typical accts. but nobody would disagree with entanglement at the center of it and entanglement + nonlocality seem to be nearly the same thing...!

Semiclassical

if one is prepared to swallow that, one gets an interesting story

but that's a hell of a price of admission

okay, back later

glS

@Semiclassical is that saying that for deBB you need to assume that the measurement choices are in some sense not uncorrelated?

ACuriousMind

19:00

@BernardoMeurer So do you have a permanent position there now? I seem to recall you were originally doing an internship

Bernardo Meurer

@ACuriousMind Yeah, they hired me part-time while I'm still in school :)

ACuriousMind

Nice!

Bernardo Meurer

b/c I didn't want to drop out

glS

@vzn not quite.. there are some subtleties here =). First, nonlocality as in impossibility of LHV models is surely important for QC, it is the no-signalling aspect that is not. Second, when I was talking about "exploiting" before I was talking about one party not being able to use in any way what happened to the other party. This is an entirely different setting that you have in QC, in which you don't really have (or at least need) spatial separations

Bernardo Meurer

Yeah :)

glS

19:02

@vzn even though to be fair, I agree that it's not clear that one can really separate these two aspects of QM

ACuriousMind

@BernardoMeurer So school's going fine, too, or is it boring you to death yet? ;P

Bernardo Meurer

@ACuriousMind Do you want the long answer or the short answer? :P

ACuriousMind

@BernardoMeurer Whichever you want to provide :D

Bernardo Meurer

@ACuriousMind I'm sick and tired of being in a CC, I can't stand it anymore. This means I need to finish IGETC (Basically a list of classes you have to take to transfer) and my major requirements ASAP. CS has a lot of requirements, meaning I'd have to be here another two years (totalling 3), so I'm changing majors. My major is not Math (obviously focusing on CS) and I'll be in some UC (Berkeley or LA) hopefully this time next year. Once there I'll declare a CS double major everything going well.

ACuriousMind

@BernardoMeurer Oh dear, I don't understand any details of this system but it sounds like a lot of pressure.

Bernardo Meurer

19:11

*My major is not Math -> My major is now Math

@ACuriousMind ^ Important difference :P

enumaris

o.o

ACuriousMind

::squints:: ...where's the difference?

enumaris

The asterisk

Bernardo Meurer

@ACuriousMind I made the same typo twice, lol

enumaris

oh he changed it

Bernardo Meurer

19:13

But yeah, it was a really tough decision, but I just want to be done with this stuff

ACuriousMind

@BernardoMeurer Apparently your subconscious really doesn't want to major in Math ;)

2

Bernardo Meurer

Don't worry, my conscious doesn't want to major in Math either

Chakrapani N Rao

19:47

Does any one have anything to say about this?

-1

Q: Can this thought experiment successfully prove that HUP is epistemological and not ontological?

Here's a thought experiment. Let's assume that we have the most cutting edge technology at our disposal. We need an electron detection sphere, as small as possible, it must detect a single electron when it hits the surface. Now assume that we introduce an electron in the sphere, as slowly as po...

quantum-mechanics classical-mechanics heisenberg-uncertainty-principle

I have tried it before, but never got a satisfying answer

Transcript for