Paul Kwait made the news by having a paper published in this week’s Nature journal for the discovery of what he calls counterfactual quantum computation.
Update 0301: Read his email posted at Cosmic Variance for his less-technical explanation of what actually happened.
Uncertain Principles has a nice description of the basic underlying principle for this form of computing, which relies of what is now known as the quantum Zeno paradox. In quantum mechanics, making a measurement forces a system to take on a particular state, which would change the outcome of the experiment. By constantly observing a system, one can force the system to essentially “freeze” into a specific state. However, one can push this idea even further and set up a (paradoxical-sounding) experiment to determine the state the system should have ended up in, even though one keeps observing the system and prevents it from evolving into such a state. In Kwiat’s work, the scientists ingeniously make use of a quantum Zeno effect to perform a quantum computation, and then set up the measurement using a second quantum Zeno effect to measure whether or not the computation was run. They can they extract the answer of the computation without even actually being sure if the computation was carried out in the first place.
Mind-boggling, I know. There are some subtleties, like that the information collected is probabilistic in nature and thus none of the answers are ever known with certainty. But with enough partial results, one can do statistical analyses to make reasonably conclusive statements.
References:
- O. Hosten, M. T. Rakher, J. T. Barreiro, N. A. Peters, P. G. Kwiat, Counterfactual quantum computation through quantum interrogation. Nature 439, 2006, 949-952.
- J. Dowling, Quantum information: To compute or not to compute? Nature 439, 2006, 919.
Urm, a Monte Carlo computer? Makes for an interesting dilemma - How do you run a Monte Carlo algorithm on a Monte Carlo machine?
Maybe you can, maybe you can’t
A quantum computer isn’t a Monte Carlo computer, by the way. The different is primarily thought to be due to quantum entanglement. (see e.g. Nielsen and Chuang for the difference between probabilistic computing and quantum computing)
Yes yes… it’s different… what I mean is the output (of a deterministic algorithm) from the quantum computer is similar to that from a Monte Carlo algorithm on a deterministic machine. (i.e. must do statistical tests for convergence, error estimates etc etc). So how will the output of a random algo look on a quantum machine?
I’ll go look up N&C to get more background on this…