Tuesday, December 02, 2014

Quantum teleportation demystified

About a week ago, NASA's Jet Propulsion Lab, where I used to work, published this news item about an advance in quantum Teleportation.  The progress reported in the story is real, but the description of quantum teleportation presented in the article is misleading at best.  Yes, there are certain compromises that you have to make when describing quantum mechanics to a lay audience, but you don't have to actually lie to them, which is what the cartoon at the beginning of the article does.

Here's a recap of the description of QT in the cartoon:

  1. Alice has a yellow photon
  2. Charlie has a pair of blue photons that are entangled
  3. Charlie sends one to Alice
  4. And one to Bob who stores it in his crystal memory bank
  5. Alice and Bobs relationship is cordial but somewhat distant
  6. As the blue photon Alice received collides with her yellow photon she measures the event and learns that the state of her photon has been teleported to Bob's
  7. Alice's measuring of the event affects Bob's far-away photon, changing its state
  8. However, Bob can't determine that his photon has been changed
  9. until Alice sends him two bits of information over an optical fiber
  10. And Bob learns his photon has changed from blue to yellow too!

There are so many problems with this description that it's hard to know where to begin, but the most egregious is the last step.  It is not true that Bob "learns" that his photon has changed state from Alice's bits.  Bob has to manipulate his photon in order to make its state the same as Alice's original photon.  The two bits that Alice sends to Bob are instructions on how Bob should manipulate the state of his photon.  In the parlance of the cartoon, Alice measures her photon and discovers that it's yellow.  She then sends a message to Bob: "Make your photon yellow."  Bob does this, and lo and behold, Bob's photon, which used to be blue, is now yellow, just like Alice's original photon.

Of course, when you add this step, the whole process seems a whole lot less mysterious and newsworthy.  In fact, it seems completely mundane.  Why even bother with the entangled photons in the first place?  Well, it's because photons are weird.  They are not either blue or yellow.  That is too simple of a model.  Photons can be blue and yellow (in the parlance of the cartoon) at the same time.  But even that doesn't really capture the truth.

A better model of a photon is a coin: there's a heads side and a tails side, but you can only ever see one or the other.  You can't look at both sides at the same time.  These photon-coins are like vampires: they don't cast reflections in mirrors.  And if you try to enlist the help of an accomplice to look at the tails side while you look at the head side, the coin vanishes.  Try as you might, you can never see both sides at once.

So you look at the heads side and it's yellow.  You turn the coin over to look at the tails side and it's red.  You turn the coin over to look at the heads side again and suddenly it has changed from yellow to blue!  You turn it over again, and the tails side, which was red, is now green.

You do this a zillion times and you discover that the coin behaves in the following way:
  1. Every time you look at the heads side, it is either blue or yellow.
  2. Every time you look at the tails side, it is either red or green.
  3. Every time you turn the coin over, the color you see bears no relationship to the color you saw on that side the last time you looked.  As far as you can tell, the color of the newly revealed side is always completely random.
  4. However, you never actually see the coin change color.  Whatever side you are looking at stays the same color until you turn it over to look at the other side.
You also have a magic coin vending machine: you put one coin in, and two coins come out.  (Let us call this Vending Machine D for Duplicator).  Each one is half the size of the original, and they come out of the machine edge-on so that you can't see either side.  But each of these smaller coins behaves exactly like the larger coins did: the heads side is always blue or yellow, the tails side is always red or green.  But here's the kicker: if you take a pair of coins that just came out of Vending Machine D and turn them both over the same way (i.e. both heads or both tails) the colors on the two coins will always be the same.  But this is true only the first time you look at any given pair of coins.  After that the coins lose their magical connection and their colors are just random when you turn the over again.

What quantum teleportation does is allow you to transfer the magic connection from a pair of half-size coins from Vending Machine D to a pair of (previously unconnected) full-size coins [1].  It uses two additional vending machines, which I will call Vending Machines A and B.  Vending Machine A is on Alice's side, and Vending Machine B is on Bob's side.

Vending machine A has two coin slots, one for a full-size coin and one for a half-size coin.  It also has a display that flashes a number between 1 and 4.  Alice inserts her coins and notes what number the machine flashes.  She then calls Bob on the phone and tells him what the number was.

Vending machine B has a single half-size coin slot, and a knob with four settings numbered, as you might expect, from 1 to 4.  Bob sets the knob to the setting corresponding to the number that Alice told him on the phone.  He then inserts his half-sized coin, which the machine spits right back out.  (Whether or not this is the "same coin" turns out to be a Very Deep Question which we will ignore here.)  Like Vending Machine D, this half-size coin comes out edge-on, so Bob can choose whether to look at the heads side or tails side first.  Whichever he chooses, the color will be the same as that on the corresponding side of the full-size coin that Alice inserted into Vending Machine A.

That is a much more accurate description of quantum teleportation, and I think it's one that a layman is capable of grasping.  So why didn't NASA describe it that way?  Well, I asked them.  I sent an email to the contact person listed on the article, and here is the response I got:
Hi Ron, I very much appreciate your feedback. I took it back to the
scientists and they explained while the photon would not be initially blue
per se, we made this choice ‹ to say that the photon is initially blue ‹
to communicate the broad idea for the public. The question of at which
moment the photon is ³changed² cannot be answered with quantum mechanics
and is too subtle a point for our purpose, which is to give the public a
general sense of this experiment.
To which I responded:
> The question of at which moment the photon is ³changed² cannot be answered with quantum mechanics 
Yes, it can. And actually, this points out another problem with your presentation: when Bob receives Alice’s bits, he uses that information to apply one of four transformations to his photon. It is Bob’s action that changes his photon’s state into the state of Alice’s original photon. And that is the *only* time that Bob’s photon changes state. 
> and is too subtle a point for our purpose, which is to give the public a general sense of this experiment.

I think you underestimate your readers, and do them (and NASA) a grave disservice by not telling them the truth, particularly since this particular non-truth makes quantum teleportation seem much more mysterious than it really is. At the very least you should explicitly say that your presentation is an oversimplification, and give your readers a pointer to where they can get a more accurate explanation. I would be happy to write up a more accurate (but still accessible) description of quantum teleportation if you don’t have the time to do it yourself.
I never got a reply to that, which is what motivated me to publish this explanation here.

It frustrates me to no end that the popular press continues to publish false and misleading information about quantum mechanics in general, and about entanglement in particular.  There is no legitimate reason for it.  So why do they do it?  Well, the description in the NASA article is certainly more dramatic and mysterious.  For starters, it sounds a lot more like "real" teleportation than the more accurate description.  The fact that Alice has to send Bob instructions on how to set the dial on Vending Machine B takes a lot of the charm out of it.  In fact, you have to think pretty deeply about the behavior of the coins in order to see why quantum teleportation is interesting at all.  After all, why can't the four settings of the machine simply be an encoding of the colors on the two sides of the coin?  To answer that question you have to get into a discussion of the Bell inequalities, but that can also be done in a way that is accessible to laymen.

So why does the popular press so consistently get this wrong?  And why does the physics community let them get away with it?  Quite simply (I conjecture) because physicists think that it's good for business for the public to perceive what they do as indistinguishable from magic.  But for scientists to allow misinformation to promulgate is hypocritical, and fraught with all manner of peril.  When physicists endorse nonsense like "Alice's measuring of the event affects Bob's far-away photon, changing its state" on the grounds that it's a harmless white lie, it makes it that much harder to combat the not-so-white lies.  In a world where rejecting science can pose existential threats, deliberately obfuscating the truth is not a trivial transgression.  The fact of the matter is that quantum mechanics is not the intractable mystery that it has been made out to be, and allowing this perception to persist for the sake of grant money is making a deal with the devil.

[1] Since I just got finished ranting about the evils of allowing scientific misinformation to promulgate I should point out that this is slightly inaccurate.  Quantum teleportation does not transfer the magic connection to a pair of previously unconnected coins, it transfers it from one member of a pair of connected coins to one previously unconnected coin.  Of course, you can do this process twice in order to produce the effect as I describe it, so I submit that this really is a white lie.  But I will let you, gentle reader, be the judge of that.


Don Geddis said...

I enjoyed your link to David Mermin's Bell inequality paper. The history was fun. But still, after all this time, I remain a little sympathetic to Einstein's objections. And I think Mermin is a little too happy to conclude that "it must all be magic that we can't understand!"

He even quickly mentions (and just as quickly dismisses) objections to "the naive classical assumption of localizability".

But I, personally, suspect that Einstein's philosophical concerns would have been answered with an understanding that the waveform is (only!) what is real, that it evolves deterministically, and that "observation" is identical to "entanglement" (as you've often said), and especially that "human observers" are no different at all from automated detectors.

Einstein -- quite rightly -- objected to the Copenhagen interpretation (= human consciousness is somehow "special", waveforms "collapse"). And to "spooky action at a distance". I think his philosophical (and physics) intuition was right, and those features don't belong in any explanation of quantum mechanics.

(Yes, Bell showed he was wrong about hidden variables, but I still think the intuition that drove him to EPR was essentially correct.)

Ron said...

If I ever find a genie in a bottle my first wish will be to go back in time and ask Feynman why he didn't think that the Von Neumann measurement model (and decoherence, of which he was also surely aware) was an adequate account of QM, and whether he still stood by his famous quip that "no one understands quantum mechanics." And then my second wish will be to bring Einstein forward in time and show him the Bell paper and the Aspect experiment. And the GPS in my iPhone :-)