Thursday, March 16, 2023

Reducing Errors by...Using Errors

 Just a quick note: our newest paper, "Leveraging hardware-control imperfections for error mitigation via generalized quantum subspace", by Yasuhiro Ohkura, Suguru Endo, Takahiko Satoh, Rodney Van Meter, Nobuyuki Yoshioka, is up on the arXiv at https://arxiv.org/abs/2303.07660!

This paper goes into a lot of detail on error mitigation techniques. Basically, in quantum error mitigation, you add errors into your quantum circuit in order to understand the effect of errors on the answer to your problem. For example, you introduce some waiting time between finishing the computation and measuring the results, to see how decoherence affects the state. Or, you add a couple of gates that, in theory, cancel out, and see how the error rate changes. These kinds of techniques have the potential to extend the range of problems that can be solved with noisy quantum computers.

This paper, joint work with NTT and Todai/RIKEN/JST (one author has three affiliations), represents an amazing amount of learning by Yasuhiro (who goes by the nickname "rum"). The math in it is captured fairly succinctly, but in fact there is a lot more underlying it, and I am deeply impressed by rum's rate of learning over the last year. He has used an astounding amount of computation time on specific IBM processors during the creation of this paper, and I would say it has paid off. I think this paper will help guide others who have read a little about error mitigation, but aren't sure how to apply it in their own context. It's the kind of practical paper that bridges theory and implementation that I consider a signature of my group.

No comments:

Post a Comment