Tuesday, February 11, 2014

Zeilinger's Newton Lecture on foundations of QM

On Sunday, Peter F. would send me a link to an obnoxious anti-quantum tirade by Adrian Kent. Except for some random uncontroversial spam about the discovery of nuclei and other things, it's nothing else than another salvo saying that quantum mechanics has to collapse because it is not classical physics – it is refusing to fulfill the obligation to say what is happening during the experiment. Well, quantum mechanics answers this question completely clearly and unambiguously when it says that it is physically invalid to ask about the measurable quantities before they're measured – they just don't have any specific properties. I decided that Adrian Kent didn't offer anything new and I've explained why folks like him are cranks too many times and it didn't make any difference whatsoever. Let me promote something more sensible instead.

In 2008, quantum information experimenter Anton Zeilinger of Vienna won the Newton Medal (U.K.). He gave a talk:

The video has 68+ minute. In the beginning, the host mentions that when Slovakia joined the EU, they would send a quantum message to the Slovak colleagues that would say – if I include the text in between the quantum lines – "Haha, you're finally back in the Habsburg Empire, you Slowakische Untermenschen!" ;-)

Zeilinger begins with a discussion of the Heisenberg Microscope and the Schrödinger's Cat. The gedanken experiments continue with those that appeared in the Bohr-Einstein debates. Einstein said that "physics is about what is out there" while Bohr would say "physics is about what we can say".

Niels Bohr is presented as a great theoretical experimenter – he wouldn't forget to draw nuts and screws etc. in an experimental apparatus. ;-)

We learn about Zeilinger's ex-adviser who would start the field of neutron interferometry in 1974-1975. Recall that neutron interferometers are also enough to prove that the "entropic gravity" is a pile of crap. Zeilinger says that he's convinced that the size of the objects that can display interference is unlimited; it is only a question of the money and experience; carbon molecules have already been shown interfering. The experimental outcome is actually prettier than the theoretical prediction (due to the Casimir interaction). It confirms Feynman's quote
Yesterday's sensation is today's calibration, and tomorrow's background.
It's because the Casimir effect used to be a sensation, now it is a background.

Zeilinger explains that in that experiment, decoherence is suppressed because the molecules cool down by the emission of radiation more quickly than the time they need to decay. He defines decoherence as a flow of the information from the system to the environment. Elaborating on this line of reasoning, Zeilinger argues that decoherence may be suppressed to a non-problem even for viruses at room temperatures! Comparably large molecules have already been seen interfering.

His folks are adding small mechanical levers. Interference should be checkable for objects with \(10^{20}\) nucleons. There is no objective border between the classical realm and the quantum realm; the quantumness depends on the experiment. A joke about quantum cars follows.

At 21:45, the second part of the talk, about entanglement, begins.

Zeilinger says that the founding paper, one by Einstein-Podolsky-Rosen, was published in PRL without a referee report! Later, PRL would start to send Einstein's papers to referees, so he would stop sending papers to that journal. ;-)

I would say that if Einstein were guaranteed to be smarter than the average referees, it would be an appropriate reaction to the humiliation by the "peer review". However, Einstein actually wasn't "generally superior" in QM issues in any sense so it was true that the peer review does help to increase the average quality of the published papers. The EPR paper would get about 3 citations per year up to the 1960s and then the anti-quantum movement really exploded, so the paper would get hundreds of citations per year ever since. Formally Einstein's most cited paper today (a really preposterous ranking).

Schrödinger would start to talk about entanglement in 1935, too.

Zeilinger says that various claims in the EPR paper are right or wrong depending on one's philosophy but the last sentence of the paper about the existence of the realist description of the experiment is definitely false.

A joke. Some people would tell Bohr to write papers more clearly. He answered: "Why should I write papers more clearly than I think?" And he had a point, Zeilinger says. At any rate, Bohr would say rather clearly that physics is about what can be said about the system, and what can be said depends on the full experimental setup. Bohr would also rather clearly say that quantum states are not real objects, they only exist in our mind. And reality may be different for different observers.

This was a Newton Lecture so Zeilinger had to say what Newton would say about this question. What he would say was written in his letter to Bentley, probably an ancestor of a carmaker. ;-) Newton found it inconceivable that inanimate matter affects inanimate matter at a distance without a messenger. Zeilinger says that Newton's claim was wrong as well. Well, I would be more ambiguous. Newton really defended locality, and he was right because entanglement isn't a "nonlocal influence". It's just a correlation. At any rate, it's silly to grade Newton in the quantum mechanics class – he had apparently no clue about the subject.

Zeilinger says that there's no mediator of the gravitational force. I disagree with that. The gravitational field, whether classical or quantum, is a mediator, and it also implies that the measurable influences are propagating at the speed of light, not the infinite speed.

In the 1960s, John Bell wrote a paper that was a useful footnote for the EPR paper. Quantum mechanics really does conflict with Einstein's local realism. There are sharper examples of the differences between local realism and quantum mechanics. Zeilinger praises Bell's paper on Bertlmann's socks. Bertlmann held the different-color socks to protest the world; he was still doing the thing at the time of the talk. In classical physics, the anticorrelation between the socks' color is trivial. Is that trivial in quantum mechanics? My answer is that the essence of the (anti)correlation in quantum mechanics is exactly the same – it follows from some correlated properties imprinted to the two subsystems in the common past, when they were in touch – and it is possible to guarantee such correlations in the sock-like way despite the fact that the individual color of the sock (its quantum counterpart) remains uncertain.

It just happened that the experiments with lasers became possibly at the same time when Bell wrote the paper. Zeilinger says it's a pity that no one has developed the school-friendly version of these laser experiments.

The term "qubit" was introduced by Schumacher: you may be dead and alive with different probability amplitudes. OK, to show that this is not a funeral joke, let me admit that his name was Benjamin Schumacher. Zeilinger would publish a picture of a nude fat Austrian porn model in Nature (to show some quantum cryptography); no one would protest because she wasn't a military symbol. ;-)

Quantum teleportation is dedicated some time. Zeilinger says that Bohr's non-realist position makes the very process of the teleportation a non-event. What you can say about the last system depends on what you've done with the middle system. For a strict Bohrian, there's just no problem here.

Entanglement at large distances. Thousands of kilometers, done. Measurements at spacelike seperations were just getting started.

At 51:00, Zeilinger is asking whether "locality" or "realism" is wrong.

This could have been a philosophical question, like the EPR original questions, but it was a damn tangible oncesince Tony Leggett's 2003 paper. Leggett's class of would-be promising "crypto-nonlocal" (CN) theories predict an inequality analogous to Bell's one but independent from that. Zeilinger falsified these non-local realist theories in 2007. He would like to decide about other classes defined by some appropriate properties.

At 58:40, Zeilinger praises the free-will theorem. After 1:00:00, he only mentions some directions in quantum computation. He picks Grover's algorithm to search in an unsorted database. You only need \(\sqrt{N}\) and not (the classical average result) \(N/2\) checks to find the person for a given telephone number with \(N\) entries.

Zeilinger predicted that quantum computers would become everyday life gadgets – which he believes especially because the decoherence problems may be overcome in his view.

He completes the story about the Moon. Einstein would ask Bohr: Do you really believe that the Moon is not there when nobody looks? It's much less well-known that Bohr would wisely reply: And can you prove the contrary to me? ;-) Physics is about what you can say. If you can't measure (or prove) something, it's not a part of facts of science.

As Zeilinger concluded in his meditations with his friend Dalai-Lama ;-), quantum mechanics is about what you can say about Nature, so the information is very important. In the beginning was the (Qu)Bit/Word. A quote by Newton who says that he is opening the secrets gradually. Finally, he encourages the young curious people.

The most famous former U.S. ambassador to Czechoslovakia in 1989-1992 (and to Ghana) died. Here in Pilsen, we would know her from the celebrations of the liberation by the U.S. army. The 1934 movie above, Bright Eyes, shows that child stars are no invention of the recent decades. RIP, Shirley Temple Black.


  1. Hi Kimmo,
    I wish you enjoy your book-launch by not forgetting to bring and imbibe a couple of bottles of Koskenkorva during it.
    I predict that your book will remain too costly for Lumo. ;-)

  2. Cynholt,

    I mostly agree with your statements here, US pigs, however, is a little broad. I also view the US as an imperial power, but it does contain many decent people (as well as many feral ones). I advocate removing all US ground forces from around the world, but building as many aircraft carriers as finances allow.

    Broadly, isn’t the situation as with Taiwan. Anyone with any sense viewing a world map must realize that the preponderance of Russia and China must strongly influence events in Ukraine and Taiwan. The bluster of US leaders in these matters is but a hangover from WWII. Speaking of which, I would think the memories of the Eastern front must still be so strong as to induce moderation in leaders on all sides in the Ukraine situation.

  3. RE: "What about looking for a way to "make the people understand that the
    climate change activists are a dangerous terrorist pseudoscientific
    organization fighting a non-existing problem and trying to control the
    rest of the mankind"?"

    I think bickering scientists can be extremely entertaining, especially when they start name-calling. But I agree wholeheartedly with this, because this is is an area the malfeasant men have found a way to scare politicians into passing ridiculous laws to squeeze more money out of thriving corporations, and dictate everything we do including what kind of light bulbs we must worship. Meanwhile, the green scare merchants are joy riding in their Lear jets.

  4. deBroglie-Bohm answers the measurement problem elegantly. The quantum system is measuring device and that being measured - they are inseparable. There is no 'collapse'.

  5. kashyap vasavadaFeb 11, 2014, 5:32:00 PM

    Great lecture. Thanks for posting the link. May be someday someone will measure interference of TRF readers!

  6. That was a great answer, Lubos. I hope the young fellow profits from it. Also to tell you thanks for the Maldacena video. I went through it once, and downloaded it for a repeat later.

  7. Lubos,

    Could you explain the difference between, "correlated properties imprinted to the two subsystems in the common past" and hidden variables?

    p.s. This is a sincere inquiry with no anti-quantum bias ;)

  8. A little-known fact connected with Shirley Temple: Billy "Bojangles" Robinson, who tap-danced with her in 1935, coined the very familiar word copacetic.

    According to Wikipedia, white people misunderstood "bojangles." They thought it meant "carefree" or "happy go lucky." Black people said that it meant "squabbler."

  9. Dear Rsala, the latter is hidden, the former isn't hidden. Their not being hidden means that they're easy to be observed.

    Moreover, the former is correlated, while the latter isn't necessarily correlated in any particular sense.

    The former exists and is a basis of all of modern physics; the latter don't exist and they are only discussed in pseudoscience.

    So the only thing they have in common is that they are "variables".

  10. I think one of the most counter-intuitive features of QM is that, while uncertainty about what we will observe is a fundamental facet of nature, we are still permitted make precise statements about classically forbidden correlations between those fundamentally uncertain observations.

  11. Hi Lubos, I think you missed the whole point this time, it happened to me firstt, but if you "read" it another way, everithing come right, like playing a Rolling Stone vinil backwards and being enlighted!

    I have a personal implementation of the idea up and running and a blog about it, so you can see some nice videos, download the exe with source code (Delphi7, object pascal, you can "read it" easily) and see it working on your own PC, or just read and comment.


    If you understand spanish, there is some more info an a video of a seminary I held last month here:


    The post will be long, but if you allows me, I will refocus the idea:

    You are an alien octopusy, your are given 8 joysticks that control a system, for instance a space ship, or a kart like in my blog, but the octopusy itself doesn't knows about the real object it will drive.

    Octopusy can drive it "cleverly" as far as you give him (or I imagine out, it doesn't matter) a way to predict where the "system" will be (in the state space of the system) after a small delta of time: If system is on state A ("state" is including the 8 joystick actual positions) and delta time pass, the system will probably be on state "B".

    The algortihm starts here: The octopusy ponder a set of possible options (push joysticks with forces (0.1, -0.5, ... , 0), a R8 vector), and for each of them, calculate 100 possible futures by simulating one delta time after another while it pushes joysticks randomly. It gives you a "spectrum" of 100 future ending points for each "option", they are yor "microstates".

    So now you need to count how many of those 100 futures are "different" for each option: round the ending "positional" coordinates of the 100 futures (some of the state components are considered "positional", like coordinates X and Y of a kart, while angular velocity is "internal"), and now count only futures with different rounded ending points.

    It is like reducing the kart's final position after a future to the tile it is in, in a grid-like space, so a tile with 5 futures ending on it count as one.

    You have it: The magic is to ask the octopusy (it is NOT a real force, it is like the force that makes you change the TV channel, so you "ask" the agent to do it) the average of all options (they where R8 vectors) averaged with the number of different futures you found for each one (normalized first).

    And the trick to make this simple algorimth to simulate any kind of inteligent behaviuor, not just what it decided to do by itself, is to play with the weights of each future in the mix.

    Each future Fi start on the same point "P0" but ends on another point "Pi". You can take any metric on the state space and define Distance(P0, Pi) to be as you wish.

    Using as the scoring of an option the Sum(Dist(P0,Pi)) for all different possible futures (instead of just the number of them) and applying the same simple algortihm to it, gives you pretty any "simple" intelligent behaviour I could think off.

    It is not an allmighty algortihm, it need more subtle refinements to be able to get close to our sophisticated behaviour, but as simple as it is, it can give you "insect like" behaviour of any kind, and using a kind of "layered" eschema you can expand it to mimic much more complex bahaviours, incluiding at the long term, human behaviour in my opinion.

    Apologies for this heavy text block!

  12. Entropic forces are NOT real forces.

    The "force" that makes you write here instead of on Facebook is not a real force, but you do write here.

    This formula "wisper" in your ear what force you should do to your joystick at any moment not to look thumb with your friends, but the real force comes from the "agent" that decides to follow those wispers and move its muscles accordingly.

    Many blogs have argued against the idea this way, but it is just misleading. The AI commanding a cuadricopter is not making any force, it is sending commands to the motors, wispering how strong each one should push to make it loop or land. Calculating the necesary force is not the same as applying it.