Friday, December 15, 2023

Annealing: Quantum and Simulated

I don't put a lot of faith in citation counts, but they are one measure of influence. Just a random tidbit on where we are...

Quantum annealing in the transverse Ising model, by Kadowaki and Nishimori, is the origin (as far as I am aware) of quantum annealing as a computational tool -- the technology used by D-Wave, one of the early quantum computing companies. It's a hugely influential idea in our field, though it is a single-purpose computational tool, useful only for finding the ground state of some system. However, it turns out that a lot of optimization problems can be mapped to this problem, so it can be wildly useful if it works well. Unfortunately, it is hard (but maybe not impossible) to do error correction on annealers, so scalability and fidelity are expected to be limited.

Kadowaki and Nishimori cite Optimization by simulated annealing, by Kirkpatrick, Gelatt and Vecchi as an inspiration. This paper is undoubtedly one of the most important papers in all of the history of computation. It probably needs no introduction, but it can be used for a plethora of optimization problems.

And so the scores -- citation counts according to Google Scholar:

  • Quantum annealing: 2,159
  • (Classical) simulating annealing: 57,085
Wow...

No comments: