A press release from NEC May 3, 2007
Posted by Geordie in Applications, QC-Related Posts, Superconducting Electronics, Superconducting Processors.trackback
According to a recent press release from NEC, the company has just demonstrated what they proudly call the “world’s first controllably coupled qubits”.
As a counter-argument to that claim, I reference this. Also I would mention that the 16-qubit processor we ran the demo on has 42 such couplers, all of which were operated simultaneously in the course of doing an actual computation.
I’m particularly interested in seeing what response–if any–this press release will get from my fellow quantum bloggers.
If anything the Aaronsons of the world will probably be very glad NEC is lying like this because it gives them hope that their fragile self-created self-contained ecosystem can be re-built just the way they like it
Go Dwave
[...] on NEC May 4th, 2007 — nextquant Geordie Rose, D-Wave’s CTO and founder posted a reply on the recent press release from the NEC Corporation concerning a publication in the current issue [...]
Geordie, I don’t see them mentioning anything about quantum computers being able to solve NP-complete problems, I think that was the main problem that Scott and others had with the D-Wave publicity.
Oh, and the NEC their press release was timed to coincide with the release of their new Science paper in which the disclose all the details of their experiment. I think a lot of people had problems with D-Wave not releasing much of the detail of your experiment, in particular the lack of information about how coherent your evolution was.
mick: yes I understand both of these points.
However focused down specifically on the coupler issue, we did publish details of the coupler and its implementation (in PRL). We didn’t do a press release or make a big deal out of it because we’ve moved on to more interesting things.
I hope that reading the NEC press release, and understanding that the NEC group knew that the claim it made in the headline was false, would give the less media-savvy of our fellow bloggers a little bit of insight into how companies use the media to push an agenda.
Communicating a message to people who are not scientists, about highly technical and difficult concepts, is very difficult. Your first point above is a case in point. The system we’re building is designed to solve NP problems. How well it does is an empirical question that hasn’t been answered yet. There is some evidence that the speed-up is in fact exponential for general types of instances classes. It’s hardly a settled issue.
The NEC release is of a different character. In the release what they are saying is clearly factually incorrect and they clearly knew this. I’m interested in seeing to what extent the blog types will take them to issue for this.
I see your point Geordie. Conveying the fine points to the media is a really difficult task and it’s easy to get it wrong or just to ignore it especially when it comes to experimental physics. There are often because there are often so many caveats that it gets more than a little nuts to try to explain them all.
I should point out that their press release is wrong for another reason other than the one that you have pointed out, well, if not wrong it is a bit misleading.
Every experiment where a two-qubit Hamiltonian coupling has been demonstrated, where single-qubit control was also possible, is a demonstration of a “tunable” two-qubit coupling. This form of quantum control allows you to simulate any two qubit interaction. Your group, and the NEC people, have demonstrated (quite well I should point out, I’m a fan of these experiments) a very specific way of generating a tunable two-qubit coupling.
I wouldn’t blame anyone for not fully explaining this detail, this would be murder to explain to your PR department unless they really really really knew what you were talking about.
[...] led to the usual fluffified popular articles; see here, here, and here for example. But to satisfy Geordie Rose’s curiosity, my hype-o-meter has not yet reached D-Wave levels, for three [...]
“The system we’re building is designed to solve NP problems.”
So is my laptop. It’s just an “unsettled” question whether it can do so efficiently
[...] Geordie incites Scott’s hypometer, but not nearly to the record setting levels of Orion times. [...]
Hi Dave: Unless you’ve got a pretty fancy hook-up I doubt your laptop processor was specifically designed for max-SAT. It’s probably just one of those crappy general purpose hacks.
Another point on this issue is that there is a ton of empirical data from running all sorts of algorithms (both heuristic and exact) on laptop-ish machines . So while the whole P ? NP thing is a mess from the theory side, in practice the evidence is pretty clear that asymptotic exponentials are what you’re going to get using what we’ve got to work with.
The same case can’t be made for AQC-ish computers, because we don’t have mountains of empirical performance data from hundreds of algorithms yet. It’s only a matter of time before we have mountains of data on hundreds of algorithms for this sort of machine–until it starts flowing in for me the question will remain open.
Hi Geordie!
// The same case can’t be made for AQC-ish computers, because we don’t have mountains of empirical performance data from hundreds of algorithms yet. //
# In dubio pro Orion?
If “the question will remain open”, why does the official ‘Introduction to Orion’ give a positive answer then?
# D-Wave’s CEO proposed a while ago: ‘If there’s any scientists out there who want to take a look at what we’re doing, they would be our guest.’
quantum bloggers?
Are there any concrete plans on that? Why not invite Dave and Scott - as the most prominent non-D-Wave
Great Geordie, the 16-qubit processor you run are very cool!