D-Wave’s black box starts to open up

One of D-Wave\\\'s chips.

D-Wave

When Lockheed Martin purchased one of D-Wave's computers, speculation ran riot in the streets. Was it going to be used to debug flight control systems? Maybe it would be used to solve design problems. Perhaps it could replace an intern and make the coffee for everyone. Or maybe someone had pulled the wool over the eyes of the senior management and bilked them of several million dollars.

The truth was more prosaic. D-Wave's computer had been installed at the University of Southern California (USC) with the objective of trying to determine what, if anything, was special about it. Since then, the results have been mostly positive for D-Wave, and the research is generating a lot of interest in Daniel Lidar's work at USC. When Lidar was invited to talk at the Centrum Voor Wiskunde en Informatica in Amsterdam, I couldn't resist jumping on the train.

Lidar is a soft-spoken theoretician who had clearly built his presentation for a mixed audience. He gave a brief introduction to the principles behind the quantum optimizer (as D-Wave is now calling it) with a nice description of simulated annealing and quantum annealing. He also gave a very short description of the D-Wave architecture, and as part of that description, he tried to answer several of the questions that generally float around. Then he got to the meat of the experiments and the results.