Can we experimentally win simple quantum pseudo-telepathy games yet?

[Strilanc]

In quantum pseudo-telepathy, isolated players with access to entangled qubits do better than classical isolated players.

I'm wondering if it's already feasible to run experiments where the simple quantum circuits are evaluated, and result in "winning the game". (Perhaps after relaxing some of the constraints, like generating the entangled qubits beforehand.)

I very theoretically and roughly know what's needed to do it: six basic types of gates, a 2 qubit computer/interferometer for Alice, and a 2 qubit computer/interferometer for Bob.

Is this within the realm of what can already be achieved? Why / why not?

 

[Strilanc]

If no one here knows the answer, perhaps you know where I should actually be asking?

 

[.Scott]

In quantum pseudo-telepathy, isolated players with access to entangled qubits do better than classical isolated players.

No they don't.

Although entangled particles demonstrate "spooky action at a distance", that action only shows itself in the statistics. Entangled particles do not provide data transmission that is faster-than-light or otherwise non-local.

It's not easy to form an English statement that describes the distinction between a statistical effect that is inconsistent with locality and an actual non-local information transfer, but that is how these entanglement properties work.

Simply having one copy of an entangled pair doesn't allow you to "tunnel" information to or from the other copy. But if Alice makes random choices in measuring her copies and Bob makes random choices in measuring his copies, when the results are compared there will be statistical evidence that Alice's choices affected Bobs results and vice versa. This evidence will appear even when Alice and Bob are far enough apart that only faster-than-light signalling could cause such statistical evidence.

The key is this: That statistical effect provides no mechanism what-so-ever for faster-than-light data transmission.

 

[meBigGuy]

I don't think anybody is saying it provides data transmission of any kind. Do you understand the game being played, and what constitutes winning? I think it is simply a way to exploit the statistical correlations that are different than classical correlations in order to get a correlation that is statistically higher than could be gotten with classical effects alone. I don't fully understand it though. (wanted someone else to slog through the equations and game-play).

In a bell experiment the correlations follow a cosine curve as opposed to a linear curve. Imagine a game that exploited that difference resulting in a higher correlation than if it were strictly classical. Effectively a complex bell experiment.

The name of the effect sucks, but they do say it is "pseudo" - i.e "fake".

It's actually a very good question.

 

[Strilanc]

No they don't.

Although entangled particles demonstrate "spooky action at a distance", that action only shows itself in the statistics. Entangled particles do not provide data transmission that is faster-than-light or otherwise non-local.

It's not easy to form an English statement that describes [...]

The distinction between communication and coordination and what is allowed by quantum mechanics is indeed very subtle.

Quantum pseudo-telepathy is a direct application of the increased coordination allowed by quantum mechanics. It does not allow you to do faster-than-light signalling. Whether or not it's non-local depends on your interpretation of QM.

I have actually gone through the details of how the game is implemented and made a simulation. The operations the players do are unitary, so they're consistent with quantum mechanics. Each operation applies only to the players' own wires ("local") so they commute and can't be used for classical communication.