Friday, August 31, 2007

Hayden, Preskill: fast information retrieval from black holes

Patrick Hayden and John Preskill, a leader in quantum computing and a winner of a bet against Stephen Hawking, use their knowledge about computer science, qubits, and error-correcting codes to say something quantitative about the retrieval of the information from the black holes. Patrick Hayden is a boy who is also the Canada Research Chair for the Physics of Information. ;-)

Well, yes, this is what I call a very good general paper about the black hole information puzzle!

These Gentlemen consider Alice and Bob. Alice wants to destroy her sex diary irreversibly. Their main conclusion is that "k" qubits of information about the initial state are retrieved as soon as "k+epsilon" qubits of the information are carried away by the Hawking radiation - which means a timescale comparable to the black hole radius. That would be really fast, indeed. Instead of acting as a perfect erased of diaries, the black hole is essentially a mirror. When the black hole size - as measured by the entropy - decreases to one-half, much more than one-half of the information is already out.

Their assumptions about the ability of black holes to deal with the quantum information at the same moment could be exaggerated and the real retrieval could be slower. And of course, the technological tricks necessary to decode the information will probably remain unrealistic forever. But their estimate could actually be parameterically exact.

Why? Hayden and Preskill also discuss the apparent violation of the no-cloning theorem, normally cured by the complementarity principle. They quantitatively show that their model how quickly the information is retrieved is barely, at the level of order-of-magnitude estimates, compatible with the no-cloning theorem: Bob probably can't quite prove that the theorem was violated and the information was doubled even though he's close to it.

Well, as you can see, I actually guess that black holes are indeed on the "edge" of being able to violate the no-cloning theorem, and because their calculation seems to confirm such a saturation, it is a great reason for me to trust them. In some sense, I feel that it is likely that there won't be any further revolutionary papers that would make the mechanism of the retrieval much more transparent than it is today.

Recently there was another preprint about the related issues:

Gary Horowitz: black hole mysteries

A light, popular review of the black hole information puzzles and the way how string theory answers at least some of the questions.


  1. Actually molt,
    almost no-one subscribes to this view
    (Preskill was one of the 3, and he said that he didn't liked the options but thought it meant 'the unitary evolution of the wave-function', and voted b.
    Without further explanation what this meant.
    I think it isn't very reasonable to state this, and then not have a single paper about it)

    Highly regarded gentlemen,
    > you are all part of the happy few who answered my question
    > a: I subscribe to the MWI that contains many parallel universes that differentiate during every 'quantum event' (meaning that there are many perhaps an infinite number of copies of everyone that inhibits earth)
    > b: I subscribe to the MWI that contains many parallel universes that split during every 'quantum event' (meaning that there are many perhaps an infinite number of copies of everyone that inhibits earth constantly being created)
    > c: I don't subscribe to a or b, because I think they are both *(probably) false
    > d: I don't prefer any of the above
    > -2 said c/d, those where counted c 1/2, d1/2,
    > -one said I can't do anything with the poll, cause this has no empirical evidence and therefore has nothing to do with physics,
    > to avoid all accusations of being biased I didn't count him,
    > -one said c with a minor change, I don't subscribe to a or b,
    > to once again avoid all acusations of being biased I didn't count him in,
    > -one said I don't like both descriptions a or b, to me mwi just means unitary evolution but I chose b,

    > The results
    > C: 18
    > D: 7
    > A: 1
    > b: 2
    > (when you search this poll you can find the full poll (with names on google I'm sure)

  2. Dear Berry, your comment is completely off-topic – this blog entry is about the information retrieval from black holes.

    It's also totally confused and uninteresting; science isn't done by polls, especially not idiotic ones like one you have sketched. Science is done by collecting and evaluating evidence and comparing it with the hypotheses and theories which are subsequently being improved or replaced by more accurate ones.