I continue to be enthused by the efforts of D-Wave Systems from Burnaby, Canada:
The Quantum Computing Company That Could
These intrepid folks have overcome the most extreme and vitriolic attacks from other members of the physics community. Rather than welcome the new approach signified by quantum annealing and adiabatic computing, the Academic Physics community issued a constant stream of negative commentary, ridicule and invective.
In their minds, the idea was pure poison, an idea to be opposed because it undermined their faith that a quantum computer should look like a classical computer! It must use gates and support the belief in quantum parallelism. Any other ideas about what constitutes computation were Verboten!
To anybody, like me, who researched quantum computing in the early days (for me starting in 1986) who was familiar with other approaches, this extreme fixity of mind always seemed as foolish as it was ignorant. Unknown, to quantum computing zealots, von Neumann himself had rejected the model of computation embodied in the von Neumann architecture. He did not think gates, or circuits based on switches, were anything but an interim idea. Physical computing is engineering, it requires creativity.
Not so for the Zealots! They had imbibed Feynman and they knew that gate based computing was the way to do things properly. Such certainty must be Bliss!
It is Bliss until it becomes a Living Hell.
Now we are at the very early stages of One Giant Backlash.
Perhaps the most manful of the early detractors to step up is Seth Lloyd. A researcher of significant accomplishment in quantum computing theory, Lloyd was recently quoted in a BBC story on the D-Wave-NASA-Google tie up as saying:
“I was probably wrong, and [Lockheed and D-Wave] were probably right”
This is in reference to his opinion that the device would not work, in spite of his own early involvement in the theoretical development of adiabatic quantum computing.
It is to be hoped that other scientists have the decency and honesty of Seth Lloyd.
The last two decades in quantum computing and quantum information sciences research have been ones of stupendous Lost Opportunity.
Very early in the development of the subject, somewhere around 1990, there was an unwarranted “closing of the mind”. We saw a select few approaches, a few among many, anointed as conventional wisdom. These select few ideas then garnered all the attention and funding dollars. The result has been an unmitigated disaster of scientific puffery of the worst kind. It has been a dreadful period of unscientific misadventure.
Thankfully, one very committed group of researchers and their investors have now shattered that world for good. Creativity is now unbound and possible.
I had gotten heartily fed up with the claims made by a few defending their “superior” knowledge and insight to distort the scientific landscape of what is “known” and what is considered “fact” and what is held as “viable” for research purposes.
This has led to a greatly distorted and warped research agenda which is overly narrow and has proven to be spectacularly unsuccessful as compared with the committed program of exploration pursued by Geordie Rose and his team at D-Wave.
Hat’s off to D-Wave!
They may well end this suffocating Dark Age which has so restricted scientific progress over the last twenty years. There is so much more that can be done in taking physics to the next level of scientific inquiry.
If we are to take those giant leaps forward we must end the Zealotry and Academic Feather-Bedding which has so bedeviled research in quantum measurement, quantum information, quantum computing and quantum cryptography.
We need a return to principles of honest scientific discourse. A properly engaged dialogue that is not caught up in the endless perpetuation of whatever dogmatic viewpoint serves the self-interest of so-called leaders in the field.
There are real advances to made in quantum theory.
Hell, in the field of self-field electrodynamics we already have a candidate new theory of quantum fields! In the work of Barut and others, you have an actual live body of referred work going back twenty-five years, which lays out a cogent and very credible basis for a new version of QED that is non-perturbative.
Why has nobody heard of this? Why are funding dollars not committed to figuring out experimental signatures of that theory and its fuller development?
I think I know why… it is the same reason D-Wave attracted such opprobrium. Those reckless Canadians at D-Wave dared to back the Dark Horse and it came home.
They were punished mercilessly for breaking ranks and trying something original!
Elsewhere, you now have a bunch of overweight, overfed and inbred Nags to put down.
It is the same in many fields of Academic physics. Money goes to whomever can form the greatest and most vocal Mystery Cult. It is not science anymore to back crap when there are better avenues available, however unpopular they may have been.
Those are the actions of a self-perpetuating Priesthood, not a community of minds who are committed to find the truth. Once you allow such attitudes to take root, as they have done, then you are left with a stupid H-Index Fueled Fund Feast:
Who can garner the most Funding Heat for the least amount of Scientific Light.
The contemporary patterns of behavior are insane!
I am hopeful that the D-Wave affair, once the true scale of it becomes public, will at least shame a few of the worst perpetrators to clean up their own scientific act.
It has been a shambles for two decades and the time for real change is now upon us.
I have my own wee challenge to get on with. When that is done, I think we can put a whole stable full of Old Nags down and fire up the glue factory.
Science and Engineering will be better for it.
All hail those Reckless Canadians from Burnaby!
They may well have saved the very soul of physics.
Been arguing on Scott Aaronson’s blog about this for what feels like months now, it is one thing if he thinks that D-Wave doesn’t have the best approach to QC, fine, it’s a free world, and making the case may very well yield some constructive criticism, but he insists that it’s not even QC. I try to get across to him that this is a lost cause. He may as well start fighting wind-mills.
Funny thing is that he clearly states that the question of the D-Wave machines’ performance will have to be settled empirically.
Of course I would I like to see other QC approaches tried as well, but why should this have any bearing on being excited about exploring what can be done now with D-Wave? I don’t understand these outspoken academics apparent fear, that somehow D-Wave’s going to “taint” their beloved field. Given the vitriol I think their fear is real, just don’t see what else would get the detractors so emotionally invested.
Scott’s not the worst of the bunch, my favorite comment there was that from an ivy league math prof who threw this out as “negative” examples:
“We’ve got D-Wave and Tesla and Virgin Galactic and Bitcoin. Each one has a lot of razzle-dazzle and has managed to pull in real money.”
Now apart from Bitcoin, which is not a single enterprise and more an ongoing experiment, each of these companies has been started by entrepreneurs who I greatly admire. What world does this man inhabit thinking that these businesses are somehow bad things?
It was interesting to me that Peter Shor (a real QC household name), who initially posted there as well, had a much more nuanced and reasonable take. He is sceptical about D-Wave’s abilities to scale up but, clearly wishes them well.
There is no doubt in my mind that eventually the “adults in the room” like Peter, taking a supportive but sceptical stance, and Daniel Lindar who embraced D-Wave early on and closely works with them, will carry the day.
The rest is just noise and a wast of intelectual energy.
Well, quantum stuff is not the whole of what has been missed here. Just try to think of a purely classical machine that does not need to either “Read” or “Write” to memory! A purely permutation machine. And yet there are so many claims in circulation, even principles like Landauers’ one [http://en.wikipedia.org/wiki/Landauer%27s_principle], all supporting certain strong ties between entropy production and computation. Noone ever wondered about the possibility of other, purely permutation based machines not requiring any act of either “reading” or “writing” in order to perform computations. And yet, the simplest example have had been given already in the famous work by Jonathan Swift, “Gulliver’s Travels”, the very famous “Engine” [http://en.wikipedia.org/wiki/The_Engine ]
Theo, I agree that the adiabatic aspect of D-Wave gets too little play. On the other hand they themselves don’t make much hay of it at this point, probably because the cooling is still too power expansive to give the machine an edge. (Unlike the distant adiabatic CMOS cousins).
But as the chips get more integrated there should only be a marginal increase for cooling power (there is, after all, only entropy increase when the final measurement occurs).
It is to my knowledge the first adiabatic foray into super-computing, and it would be great if it could eventually contribute to the slowdown of the ever increasing power hunger of data centers.
lambda, you write “Rather than welcome the new approach signified by quantum annealing and adiabatic computing, the Academic Physics community issued a constant stream of negative commentary, ridicule and invective.” But this very community you accuse of issuing the stream of negative commentary, not D-wave, came up with the idea of adiabatic quantum computing , and has done a tremendous amount of work on it. It has also done a fair amount of work on computing via cooling to the ground state of a Hamiltonian, which could be one thing that is meant by “quantum annealing”. (There are various things that go by that name… there is even a classical computational strategy which solves problems by classically simulating the cooling of a quantum system toward its ground state.) So I think you are greatly underestimating the work of the academic quantum computing community… without which D-wave would not exist.
Reasonable Riposte, Howard. Tell me this. Was the original paper on Adiabatic Computing published, and where?
I mean this one:
http://arxiv.org/abs/quant-ph/0001106
Hi Kingsley—
No idea whether it was published other than on the arxiv. But that is not very relevant to my point that these ideas came, in substantial part, out of the academic physics community. As I’m sure you know, Farhi and Goldstone are famous physicists, tenured at MIT. Sipser is at MIT in math, Gutmann at Northeastern in math. I read their paper soon after it first came out, and I’m sure they could have found somewhere to publish it if they’d wanted to. Again, I don’t know whether or not this paper, or some version of it, was published elsewhere or not because the arxiv pretty much suffices to get people’s attention in quantum computing, if your work is good enough, and that paper was recognized as seminal almost immediately. Those guys are at a point in their career where they don’t have such a great need to publish in journals if they don’t want to, though I suppose it still helps with grant renewal.
I do think that elements of the academic physics community in the US have been slow to catch on to the importance of quantum computing and quantum information…primarily in mid- to lower- level departments, though it’s very dependent on who is already there. I think that is changing, although the receptivity to experimentalists is still probably greater than to “high theory” and especially interdisciplinary theory that is seen as “not really physics” (quantum algorithm theory, etc…). Europe seems to me to have been faster to embrace QC/QI. Caltech and MIT were way out front on this in the US, though, and some other institutions (e.g. USC, U Lousiana Baton Rouge) have realized that it can be a good strategic focus, and institutions with strong theoretical/experimental groups in areas that turn out to be related to QC implementation have also turned fairly hospitable, though again I’m not sure whether this spills over a little bit into receptivity to the more abstract side of QC/QI theory. I do think there were and still are (to a somewhat lesser extent) serious difficulties in the US for QC/QI theorists with physics degrees who don’t also specialize in some more traditional area of physics, especially if they are looking for an academic job…
Publication in the journal Science of an article by these folks plus some others, in the year following the original preprint you cite, including simulations of quantum adiabatic computation applied to NP-complete problems, also hardly suggests resistance by the academic establishment:
http://www.sciencemag.org/content/292/5516/472.short
Howard, there is no doubt D-Wave wouldn’t exist without the academic research in this area, after all Geordie founded this right out of university after he finished his PhD.
Nevertheless, it is astounding how much headwind this company now faces from some academic circles. Having not followed this story from the beginning, I tried to trace the source of this controversy, and it all seems to go back to an early ‘sale-pitch’ demo that thoroughly rubbed some QIS folks, especially Scott, the wrong way.
Hi Howard,
Appreciate your response and link to where it did get published (eventually). The purpose of my stirring activity is to flush out such discussion. I accept the points you make concerning the contribution from academic research. However, the contention of my remarks has been to highlight what seems to me to be some largely one-way traffic. Academics seem quick to dismiss non-Academic work and eager to claim credit for any advance outside of Academia. This is natural and fully understandable but comprehensively nutty. Over 75% of PhD graduates work outside of the Academic system. It is dumb to contend, as some in the Ivory Tower do, that they are the only source of true insight. This is the thing which annoys folks outside of the Academic system. I have seen it in Physics, Mathematics and Finance. When you make original contributions to more than one field and see them ignored you tend to get a bet jaded with the typical line which emanates from an Academic publicity machine.
If folks in Academia knew where the true state of the art is they might not come over so arrogant. The truth is that they are behind, sometimes by a long way.
Kingsley
Hi Kingsley—-
Worthwhile discussion, to be sure. I’m just not quite so sure that the academic side is as arrogant and quick to dismiss non-academic work as you say, but you may have a point. Regarding quantum information and computation, I do find it quite significant how large of a contribution was made by “non-academic”, that is to say, industry and government, researchers in the early days of the subject. But a balanced interchange between academic and industry researchers seems to have been crucial. There were the first inklings of quantum computing, primarily Feynman and (so I have heard, but have not seen the sources) also Yuri Man’in, coming from famous academic physicists, also Paul Benioff at Argonne National Laboratory in the US (govt or managed-under-contract-for-government lab). For quantum crypto, not sure where Wiesner was, but the further development had Charles Bennett as a key player, who was in industry, with IBM, collaborating with Gilles Brassard, at a Canadian University. In QC, a key insight on just how interference can be used to do things faster than classically was from Deutsch in Oxford, further developments involved both academics (Jozsa, Vazirani, Simon) and people working for industry (Peter Shor at what was then known as AT&T Bell Labs). For error-correcting codes, we have again the key insights coming from all three areas (academia, government, industry) (Steane at Oxford, Shor and Calderbank at AT&T, Rains who was at AT&T for a job interview that turned into a research session, Knill and Laflamme at Los Alamos National Laboratory (essentially a government lab, although run by Univ. of California under contract to the government at the time) .
One of the lessons that I think the history of quantum computing and quantum crypto holds for us is that maintaining a diversity of types of research institution, and modes of support for research, is desirable. In light of this, demise or downsizing of relatively research-focused corporate institutions is worrisome, although I don’t have a clear picture that this is a quantitative trend… I guess I have a sense, perhaps wrong-headed, that AT&T and Lucent have downsized their more theoretically-oriented groups, at least in quantum matters and theoretical computer science oriented stuff, and I may be misinformed. IBM maintains a small, but very strong group in quantum stuff, as far as I know. On the other hand Google seems to try to balance mission-focused work with the opportunity for “blue-sky” thinking with their one-day-a-week-to-do-what-you-want policy (if that’s still in place), and as far as I know Bell Labs always combined an ethos of willingness to work on practical problems at hand, with hospitality to less application-focused work.
Anyway, just some thoughts. I do think your own early work on Bayesian estimation of quantum states may have not gotten the attention it should have, and that may be in part because you went into industry and worked on other topics and so were not at many of the early conferences and workshops on QC. That may be indeed a hazard of a truly interdisciplinary career… it takes effort to maintain contact with the “community of obsessives” interested in a particular topic and sometimes it is better not to obsess so much on one thing. Your work was the earliest that I’m aware of explicitly using Bayesian estimation of quantum states (I could easily be unaware of e.g. some early Russian work on state estimation…) but quantum state estimation hasn’t really been a research focus of mine, so I don’t know how aware that subcommunity is of your work. As you probably know, the formulas you derived for integrals over all of state space were crucial to my early work on information-disturbance relations, and cited in my thesis and in the arxiv preprint that later came out of it,
http://arxiv.org/abs/quant-ph/0205155
but I was slow to get that paper on the arxiv and for complex reasons, withdrew it from publication (was in discussions with someone about possible joint research that looked like it would lead to a closed form expression for the tradeoff, but that never quite got finished). So it got less of a boost than it might have from my citations (not that I flatter myself that they would have been a big deal…).
Anyway, as I say, just some thoughts. I have always been struck by how crucial to the development of this field was the synergy between academic research institutions and more mission-focused governmental and industry ones.
Howard
Hi Howard,
Thanks for those thoughtful insights and the deep read on history. It would be great to have you write a piece for this blog or cross-copy one of your posts.
The entire point of my stirring is to elicit some deeper two-way traffic between industry and academe and between history and current trends.
I happen (rightly or wrongly) to think that large parts of a putative new theory (really a re-made and renovated old theory) are already in print.
However, I do not believe I know precisely where it is or which bits are most promising. If it is any consolation to the defense of academe, I think that my quest for such new theory is largely an archaeological dig.
The benefit of having departed the cut and thrust of Ultimate H-Index Fighting is that I have other means of support and so I view literature excavation as more like art collection than work. Nonetheless, I do have a pecuniary interest. If I can find the new theory among the flotsam and jetsam of peer-reviewed literature then I have a good mind to bottle it and sell it down the river to industry.
Despite my odd outburst of negativity I do rate this a fine Magical Mystery Tour. After all, a simple quantum theory would surely sell a lot of Corn Flakes.
A free wormhole to meet Mickey Mouse in Multiversal Wonderland in each pack!
Best, Kingsley