Defeating Heisenberg's Uncertainty Principle

eldavojohn writes "As we strive closer and closer to quantum computing, physics may need to be improved. A paper released in Nature Physics suggests that the limit defined by Heisenberg's Uncertainty Principle can be beaten with quantum memory. From the article, 'The cadre of scientists behind the current paper realized that, by using the process of entanglement, it would be possible to essentially use two particles to figure out the complete state of one. They might even be able to measure incompatible variables like position and momentum. The measurements might not be perfectly precise, but the process could allow them to beat the limit of the uncertainty principle.' Will we find out that Heisenberg was shortsighted in limiting the power of quantum physics or will the scientists be surprised to find that such a theoretical scenario — once conducted — performs unexpectedly in Heisenberg's favor?"

Comments

Bad choice of names?

[Krokant]

For those interested, the preprint of the Nature article can be found at: http://arxiv.org/abs/0909.0950

However, I don't really see what the fuzz is about. What they are in fact demonstrating is a relationship between conditional von Neumann entropies, which they claim is a measure of "uncertainty" (it is in a specific meaning of the word "uncertainty"). However, there is a difference between von Neumann entropy and the variance of a physical observable as used in the Heisenberg uncertainty principle. On the other hand, if you label a physical property such as entropy "uncertainty" and demonstrate a relationship between those entropies, then you can indeed call that an "uncertainty relation" but that's just a cheap way of attracting attention.

Also, I am not sure if it is possible to obtain the Heisenberg uncertainty relation from their equation. I would expect that, for example by entering pure, disentangled states in their equation, that Heisenberg should be recoverable (because of course, Heisenberg also applies to pure states). I don't immediately see how that can happen since the von Neumann entropy for a pure state is zero. Perhaps I am just missing something and perhaps my QM is a bit rusty :).

Re:EPR

[DMiax]

Yes, it was. The point being that after you do any measure your state is no more correlated and the second measure does not project the state of the first.

I read the arxiv version of the paper (later I will have to go down to the library to get the journal one) and it seems that they simply reframe a lot of common knowledge in a different terminology. It is not like they show incompatible observables measured at the same time. Measuring position and momentum of different particles is not a problem since they do commute.

P.S. The article defines Paul Dirac as "another physicist". Just look at his page on Wikipedia for Landau's sake.

Re:Afty0r

[pclminion]

For example the uncertainty in momentum multiplied by the uncertainty in position for a particle must be greater than or equal to h/4pi. Breaking that limit would break Heisenberg, even if the results still weren't totally totally certain, accurate and precise.

Breaking that limit would break the mathematics of quantum physics, not just Heisenberg. The momentum and position wavefunctions are simply the Fourier transforms of each other. If position is precisely known, then the position function is an impulse, and the momentum function must be a wave that extends throughout all space. This is simply the nature of the Fourier transform. If the uncertainty relation between momentum and position did not hold, then it would mean that the momentum and position wavefunctions are NOT the Fourier transforms of each other, and that would mean that all of quantum mechanics is wrong.

What's been demonstrated here is, very clearly, not that.