University of Cambridge > > CQIF Seminar > Quantum annealing on Ising spin glasses

Quantum annealing on Ising spin glasses

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Sergii Strelchuk.

While a universal quantum computing is not yet within our reach, quantum computing technology has matured to the point where certain quantum algorithms and simulators can be realized in laboratories. One example of this is quantum annealing which is realized in D-Wave One and Two. In this talk, we review some aspects of the recent development of quantum annealing with focus on D-Wave One and Two. First we discuss various models which potentially could explain the behaviour of D-Wave (DW) machines and we demonstrate that the statistics of DW are well explained using path-integral quantum Monte Carlo quantum annealing (SQA). Next, we look at the scaling of DW and compare this to simulated thermal annealing (SA). For the chosen problems, we show that SA is always superior to DW when looking at the scaling, and that the two approaches are about equally fast if one looks at total time to solution. That SA should scale better than SQA /DW is in contrast to a theory formulated in 2002 where it was shown that SQA is more efficient than SA in finding low-energy states. To resolve this discrepancy we revisit the work by Santoro. Studying thousands of problems, we show the better scaling report by Santoro et al. is a result of certain assumptions in the derivation of SQA . The consequence of these assumptions is that the theory does not apply to physical systems, and in particular not to DW. To further support the claim that SQA should scale worse than SA, we look at the distributions of time-to-solution. We demonstrate that the SQA distributions are dominated by long fat tails. In turn, this means that the mean time-to-solution for a large problem set is expected to be worse for SQA than SA. Finally, we discuss some of the open questions.

This talk is part of the CQIF Seminar series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.


© 2006-2024, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity