Teleport, tunnel, adiabat—but one way or another, get there
As a general rule, I don’t post workshop announcements on this blog: if I did it for one, I’d have to do it for all, etc. etc. But I’ve decided that an exception can be made, if the requesting party has won a bet against Stephen Hawking. And so it is that I, on behalf of John Preskill, hereby encourage you to attend the Quantum Information Science Workshop in Vienna, VA, from April 23-25, which has been hastily called in response to the report A Federal Vision for Quantum Information Science. The whole quantum information community is invited, but the deadline for the workshop hotel rate is today! The future of our entire field will be decided at this workshop:
- Should more quantum algorithms be discovered, or not?
- Is battling decoherence important, or unimportant?
- Are interdisciplinary insights needed from CS, physics, and other fields, or will a single discipline suffice?
If you’re as hungry for the answers as I am, you won’t want to miss this.
Follow
Comment #1 April 7th, 2009 at 1:05 am
Is admission free? Can anybody go? What is the deadline for registering?
Comment #2 April 7th, 2009 at 1:31 am
rr: John tells me everyone is welcome. I’ll defer to him or others for the other two questions.
Comment #3 April 7th, 2009 at 6:37 am
Those are some pretty important questions ;). I, too, am eager to find out the answers!
Comment #4 April 7th, 2009 at 6:49 am
Although I won’t be able to attend, I would like to report from the 50th ENC Meeting in Asilomar. that large-scale quantum spin simulations are being used by almost every research group. Good! Yet surprisingly—and regrettably—there was very little representation at the ENC of the QIS community (except that Anton Zeilinger gave a plenary talk invited by Alex Pines).
To appreciate why this disconnect between fundamental QIS theory and practical QIS application is unfortunate, it is helpful to read the sobering Wikipedia page on the “AI Winter”. The AI community has made many attempts to “purify” their discipline, for example, by deciding that “AI is all about formal reasoning from expert knowledge.” By and large, these attempts at purification have been harmful, and the result has been the AI Winter.
Is there a “QIS Winter”? Surely, many young researchers would say so.
How can the QIS Winter be shortened? That’s a great topic for the Workshop!
Comment #5 April 7th, 2009 at 7:27 am
John Sidles says:
(1)Is there a “QIS Winter”? Surely, many young researchers would say so.
(2)How can the QIS Winter be shortened? That’s a great topic for the Workshop!
Excellent questions John!
Comment #6 April 7th, 2009 at 9:26 am
…is more federal funding for QIS needed, or less?
Comment #7 April 7th, 2009 at 9:50 am
John,
I guess you think that a QIS Winter is made more likely since the list of invited speakers seems weighted toward “old school QIS”.
I don’t know if that’s the case, but I assume that the answers to the questions Scott poses will be fully addressed in the “Quantum Complexity and Fundamental Physics” presentation by one of the QIS old-timers?
Comment #8 April 7th, 2009 at 10:28 am
Michael Says: I guess you think that a QIS Winter is made more likely since the list of invited speakers seems weighted toward “old school QIS”.
My opinion is rather “healthy ecosystems are almost invariably diverse ecosystems” … as discussed at the end of an earlier Shtetl-Optimized thread.
Scent_of_Violets then raised a good point about the fine line between “a chaos of competition for scarce resources and monocultural group-think.”
This is indeed the Scylla and Charybdis between which the Quantum Information Science Workshop will seek to guide us.
I definitely do *not* pretend to have any answers better than rational debate, wise compromises, respect for traditional scientific norms, willingness to learn from history, and optimism for the future.
You know … the usual outcome of any workshop! 🙂
Comment #9 April 7th, 2009 at 11:03 am
Scott you forgot my favorite quantum question: Should we invade Canada to take over D-wave?
Comment #10 April 7th, 2009 at 11:12 am
Dave: What, launch another expensive invasion, only to find out once we get there that the much-hyped Weapons of Mass Sudoku don’t exist?
Comment #11 April 7th, 2009 at 2:20 pm
Can you suggest any questions that are not only interesting, but whose answers are less than obvious?
Comment #12 April 7th, 2009 at 3:51 pm
Greg: QMA=QCMA?
Comment #13 April 7th, 2009 at 4:46 pm
I meant questions that will be “decided” at the workshop. Or is that what you also meant? If you folks decide whether QMA = QCMA, that’s great!
Comment #14 April 7th, 2009 at 5:12 pm
I don’t know! I guess I’ll find out when I’m there.
Comment #15 April 7th, 2009 at 6:37 pm
Although I’m just an undergrad math/physics major I look at this as a great opportunity I’d be foolish to pass up. I’ve been independently studying quantum information theory for the past two years and cannot wait to attend this workshop. I’m just glad I checked other websites for hotel fees because I found a room there for 79$ a night. While it won’t be as luxurious, I’m sure it will be good enough. The only thing I have to worry about now is how to get from BWI to the hotel and back without spending an arm and a leg on cab fare.
I’ve been a lurker here for a while and I hope to see you at the workshop and am looking forward to your talk.
Comment #16 April 7th, 2009 at 7:20 pm
According to today’s David Brooks NYT column on moral reasoning, a successful workshop will be marked by shared feelings of “awe, transcendence, patriotism, joy and self-sacrifice.”
How can we get to these feelings, starting from Scott’s suggestion of QMA=QCMA? The motivation for this starting point being, the shortest posts by Scott are the often *most* rewarding to take seriously. 🙂
Hmmm … let’s see … if we start with a recent talk by Schuch and Verstraete Interacting electrons, density functional theory, and quantum Merlin Arthur … and we add in some geometric insights from Terry Tao’s recent post on Gromov’s nonsqueezing theorem in symplectic geometry … then we can begin to see avenues for gaining a deeper understanding, via informatic geometry, of which quantum systems can/cannot be observed (via classical information channels), and can/cannot be simulated (with classical computing resources) … and we begin to conceive that perhaps all of system-level biology may belong to the can-observe/can-simulate quantum class.
Yep, that definitely does the job for me! 🙂
Comment #17 April 7th, 2009 at 10:52 pm
This may be a really stupid question but what are the odds that the release of this document and the forming of the subcommittee will lead to increased federal funding for QIS?
Comment #18 April 7th, 2009 at 11:06 pm
Looking at the first sentence under “Conclusion” in the Federal report, it sounds like concrete goals and priorities are in demand more than anything else.
Comment #19 April 8th, 2009 at 4:47 am
My opinion about this workshop, if anybody is interested
Comment #20 April 8th, 2009 at 7:47 am
rrtucci,
You say – “The list of speakers includes no business interest. No entrepreneurs, no investors, no CEOs , no real business people, no software developers (of either closed or open software). How does Mr. Marburger expect to develop a new industry if he excludes business interests from day one?”
A certain very ambitious company aside, who would be on such a list? And isn’t there a risk in generating a lot of enthusiasm and investment for such a premature technology?
Comment #21 April 8th, 2009 at 9:19 am
rrtucci: Well, I was hoping to keep it from the masses, but you’ve indeed cracked the sinister conspiracy of the tenured (or in my case, not-yet-tenured) quantum computing “experts.”
(1) Yes, the workshop is a closely-guarded secret. The organizers begged me and other bloggers to publicize it merely to create misdirection. When the rubes arrive at the Tysons Corner Marriott, they’ll discover that the real workshop isn’t there at all, but in Dick Cheney’s former underground bunker … no, wait, I shouldn’t have said that…
(2) I agree, the presence of so many technical talks at a gathering of researchers is suspicious. I mean, a bunch of academics sitting through each other’s presentations, staring at PowerPoint slides, with breaks only for coffee and little honeydew slices? That’s just not the kind of thing these people would waste their time on. Don’t they, by definition, already know everything about each other’s work?
(3) You’re right, the lack of industry participation (not counting Charlie Bennett and Barbara Terhal from IBM and Mike Freedman from Microsoft) raises red flags all over the place. Where are the all mom-and-pop quantum computer stores and the roadside factoring stands? As brisk a business as they’re doing already, don’t they want even more money from the government? (It seems to be in vogue nowadays.)
Given all this, it’s only with profound trepidation that I leave the link to your blog standing.
Comment #22 April 8th, 2009 at 9:35 am
To respond to rtucci’s and KaoriBlue’s comments, the Wikipedia page on quantum chemistry and solid state physics software provides links to numerous commercial and open-source simulation codes. These quantum simulation codes increasingly are the foundation for enterprises that seek to press against the limits of size, speed, sensitivity, and power efficiency. Which is to say, pretty much all 21st century high-technology enterprises.
The accuracy, efficiency,and capability of these quantum simulation codes has exponentially increased over the past several decades. And there is every likelihood that this exponentiating pace of progress will continue, or even accelerate, for the next two decades at least.
Ideas from fundamental research in QIS are beginning to impact all of these simulation codes; this was a major topic of conversation at the 50th ENC, where most research groups now rely heavily on simulation codes to plan and interpret experiments.
The specific QIS objective toward which ENC code developers are aiming is “QIS crossover”. As sample sizes in magnetic resonance steadily decrease, and as simulation sizes steadily increase, there eventually comes a time at which researchers can simulate the quantum dynamics of every spin in the sample. For biological samples, it appears that QIS crossover will occur for the spectroscopy and imaging of biomolecular-scale systems having on the order of 103 spins.
That QIS crossover is an important milestone has been appreciated since the days of von Neuman and Weiner. That’s why it was terrific to see names like Anthony Leggett and Alán Aspuru-Guzik on the invited speaker list; these researchers appreciate the unbounded scope and scale of this near-term QIS opportunity.
Comment #23 April 8th, 2009 at 10:58 am
As brisk a business as they’re doing already, don’t they want even more money from the government?
I asked about this at the local Qubits-R-Us. They had heard about the workshop at the last minute, but they demurred that they don’t need the government’s help anyway. They would rather succeed without a crutch from the taxpayers.
(Interestingly, the small bookstand in this store carried not only “Atlas Shrugged”, but also “Why Government Doesn’t Work”, by Harry Browne and “The Ethics of Liberty” by Murray Rothbard. Okay, they also had QCQI by Nielsen and Chuang, or at least an empty space for it. They plan to replace that one with an original Qubits-R-Us audiobook.)
Comment #24 April 8th, 2009 at 12:39 pm
Scott asks: Where are the all mom-and-pop quantum computer stores and the roadside factoring stands?
Scott, that is a very cogent question, which can be given a serious and wholly non-satiric answer!
During the electronics-based information-theoretic revolution of the 1950s the public could read in Time Magazine:
It is striking that today’s public is similarly baffled and intimidated—and yet optimistic—about the term “quantum information” as it was by “electronics” in the 1950s.
Isn’t today’s promise of QIS is even greater in scope and scale than the promise of electronics? What is required for QIS to become an engine of enterprise for our decade?
These are serious topics for the workship.
Comment #25 April 8th, 2009 at 12:58 pm
John, one striking difference between electronics in the 1950’s and quantum information science currently is in number of applications that John Q Public would care about. It saddens me to say it, but until we can show that a quantum computer will help the average citizen perform some mundane day-to-day task more quickly or cheaply, we won’t see the excitement that electronics did in the 1950s.
I’m not trying to say that breaking RSA or simulating quantum systems isn’t something that is an insanely cool breakthrough (to me and you), but I’m willing to bet that your average Joe couldn’t care less and THAT is a problem imho.
Comment #26 April 8th, 2009 at 2:41 pm
Michael, your post suggests an interesting topic for debate: Resolved: the central topic of quantum information science is quantum computation.
Recognizing that there are cogent arguments on both sides of the question, recent advances in QIS make arguing the negative of that proposition seem easier. Not for any negative reason, but for the “strictly positive” reason that the increasing scope and power of modern QIS is beginning to span application domains, beyond quantum computation, that the public definitely does care about.
Taking IBM as an example, the research of Charles Bennett and Barbara Terhal represents “classic QIS”, while Dan Rugar’s and John Mamin’s spin microscopy research represents “crossover QIS”.
Scott, with MIT now hiring Christian Degen out of Dan Rugar’s IBM lab (as I hear), your university has positioned itself perfectly to pursue *both* classic and crossover QIS … this was a strategically solid hire by MIT!
Comment #27 April 8th, 2009 at 3:14 pm
John, you make an excellent point and I must concede that yes, there are numerous applications that QIS will/can provide that would be useful to everyone. However, off the top of my head I can think of only quantum cryptography as something that has received enough publicity for the public to be both aware of and , in the public’s opinion, in need of.
Now, I know there are classical encryption schemes that will require a relatively large number of qubits to solve but I doubt that this fact is helping people with very important data to protect sleep easier at night.
I’m not trying to poohpooh the efforts of QIS to make itself applicable to numerous varying fields, in fact I celibrate this fact. What I’m trying to say I guess is that the average citizen is oblivious to these applicabilities and I have to admit that before you posted on this I was unaware of several of them as well. I’m hopeful that QIS will become a more public, more “advertised” domain of science in the near future.
One problem that I’m distinctly made aware of as an undergraduate trying to get into the field is the sparsity of people, even at my university, that have any knowledge about QIS. Now, I’m biased on this I’m sure because I go to a rather small school that lacks any graduate program, however I still find it disconcerting that on my campus there is only one person I’ve come into contact with that has any higher than popular science level knowledge of QIS. Unfortunately however, that person is a Mathematics PhD who only knows about QIS tangentially through her previous work on quantum control many years ago.
IMHO, QIS is not a field so difficult that its research should be confined to the big name schools or even to graduate school at large. Most of the texts on QIS I’ve gone through only assume a knowledge of linear algebra and a certain level of mathematical maturity. I don’t see any reason why this subject could not be studied by undergraduates. Furthermore, I think that its study by undergraduates would help to provide another application of linear algebra (reinforcing what was learned there) as well as an approach to QM that isn’t nearly as esoteric as some QM texts seem to be.
Comment #28 April 8th, 2009 at 6:15 pm
Michael Luvaul Says: … the average citizen is oblivious to the applicabilities [of QIS] … QIS is not a field so difficult that … that this subject could not be studied by undergraduates.
MIchael, that was a fine, thoughtful post! I would like to respond specifically to the above two points, which have come up (for me) at least three times in the last month: (1) in the “classic QIS” context of the SQuInT Annual Meeting, (2) in the biomedical context of a visit to Rob Tycko’s lab on the NIH Campus, and (3) in the “crossover QIS” context of the 50th ENC Annual Meeting.
The same QIS material was discussed at all three venues, from three very different (yet closely linked) points of view. At the NIH, the scientific focus was on the structure of the amyloid proteins of human neuronal synapses, whose misfolding has been implicated in Alzheimer’s disease. The devastating effects of Alzheimer disease on patients and their families is something the public understands!
Magnetic resonance is of the few tools available for studying these molecular-level conformational disorders, and at the 50th ENC we discussed how to quantum-simulate the (highly complex) spectroscopy of these proteins, and also how to directly image the conformation of individual molecules by MRFM.
The classic QIS relevance is that existing spin-simulation tools (like SIMPSON and SPINEVOLUTION), in their present form, cannot accommodate the dozens to hundreds of spins that experimenters would like to simulate. So at SQuInT the emphasis was on how fundamental insights from classic QIS are essential to the next generation of quantum simulation tools like SIMPSON and SPINEVOLUTION … tools that many people (including me, right now!) are now coding-up at a furious rate. Without the classic QIS insights of the last decade, we could make zero progress on this medically vital task.
Michael, with regard to your second excellent point that “QIS is not a field so difficult that … that this subject could not be studied by undergraduates” you are IMHO exactly right! And the reason is given by a quote from Bob Geroch that appears on one of the slides of my talk (slide 25):
A program that takes Geroch’s principle to heart is the (aptly named) Perimenter Scholars International (PSI), which in ten months gives students both a master’s degree and (more important) a solid grounding in quantum information theory. Michael, the PSI program is definitely one that you should look into.
Just to say something controversial, if I were given choice between the affirmative and the negative of the following debate topic: Resolved: the primary focus of quantum information science in coming years will be team-building … I would choose to argue the affirmative.
The team-building focus of next-generation QIS will be driven primarily by the problem-solving demands of the job-market. Example: the NIH too recruited Christian Degen, but narrowly lost out to MIT. Both institutions appreciated the unbounded scope of the enterprises that are now being opened by crossover QIS capabilities.
As the math/science/engineering capabilities of QIS continue to increase, it seems likely (to me) that many more such QIS career opportunities will open up. This is good news for young QIS researchers.
Comment #29 April 8th, 2009 at 6:30 pm
Another difference between “quantum information” now, and “electronics” in the 1950s: As I recall from the 50’s, there wasn’t much of a question about if electronics worked or not.
Comment #30 April 8th, 2009 at 6:42 pm
Thanks for that response John. As for the quote from Geroich, I believe I read that a while ago in his text Mathematical Physics although I could have seen it quoted elsewhere. (That is a terrific text by the way)
Also, the application to medical problems such as Alzheimer’s is obviously something the public would care greatly about. However, I didn’t know about it until I read your post and I try my best to keep abreast of the field as much as possible (while not flunking my courses of course). So I don’t think its an issue of the problems QIS can solve not being important or understandable by the public. I’d say the problem is more so one of the public KNOWING that these problems have solutions that may be found in QIS.
As far as team building is concerned, I couldn’t agree more. Something associated with that issue should come up, and has been addressed in the past (but I feel not solved) is not only team building across institutions but team building within institutions. Has anyone yet found a good solution to the problem of what department in which a QIS class belongs (again, keeping in mind that not every institution can afford or is willing to create a separate department)?
Thank you for suggesting the PSI program as I had not heard of it previously. Although, I’m assuming it is in Canada, and my wife hates the cold, so I doubt she’d share in my thanking you for the suggestion.
Comment #31 April 8th, 2009 at 9:30 pm
Michael and Raoul: I agree that electronics in the 50s is not the best analogy, among other reasons because we’re still missing the quantum analogue of the transistor. (Not that the transistor didn’t itself require quantum mechanics…) The analogy I like is to Babbage’s Analytical Engine—with the caveat that we already know much more about the theory of quantum computers than Babbage knew about the theory of classical computers.
Comment #32 April 8th, 2009 at 10:06 pm
I would have to agree that this is much more analogous to the situation with Babbage than 50s electronics.
I also would like to point out that not only do we know more about quantum theory than Babbage did about classical computation, but also we know more about how classical computation has been applied which gives us a great deal of insight into how quantum computation might be applied.
However, I think working in an “old” theory (old here meaning 80 years old as opposed to brand new) can also have draw backs in that people get stuck thinking in the established paradigm. There are some signs out there that this may not be a problem here however. For instance, I do find it promising that there are other models of quantum computation besides just the gate model floating around. Even if it turns out that the gate model is the most efficient in the end, I am glad to see that many people are thinking outside of that model and trying every possible way to take advantage of quantum theory to contribute to the state of QIS. (caveat: I’m not a D-wave mole surreptitiously touting adiabatic QC, I promise)
Comment #33 April 9th, 2009 at 11:05 am
Hmmm … I think that deeper lessons for QIS can be extracted from both the 1950s electronics boom *and* from the history of classical computation technologies.
After all, just because Time Magazine 1957 special issue The New Age was all about electronics, doesn’t mean that the “new age” economic boom of the 1950s was all about electronics.
If we read a little more deeply into the technological history of the 1950s, we discover that the 1957 economic boom was in large measure catalyzed by John von Neumann’s 1953 Strategic Missile Evaluation Committee (SMEC), which included as members both Simon Ramo and Dean Wooldridge—who were both the subjects of that Time Magazine special issue, and in large measure, its authors.
Regrettably, the SMEC Report itself remains classified (although no doubt some attendees at the QIS Workshop will be familiar with it). However, we can learn a great deal about SMEC Report’s contents by studying the (unclassified) umbral documents of the era, for example Booton and Ramo’s 1984 IEEE article The development of systems engineering, and of course, the massive literature on the Space Program whose creation (and technological roadmap) were catalyzed by SMEC.
Of course, modern computers were being created at the same time … by the same group of people (most especially von Neumann) … using similar social pathways and technology development roadmaps.
For the above reasons, my own preference would be to argue the affirmative of this proposition:
We can ask: will the Vienna Workshop on Quantum Information Science (QIS) emphasize Babbage-style fundamental QIS research, or will it emphasize von Neumann-style quantum system engineering?
In the posts so far on Shtetl Optimized, there seems to be a strong preference for a Babbage-style workshop. Whether this is what the sponsors of QIS research are looking for, is a very interesting (and contentious) question indeed.
My own common-sense opinion—which is solidly grounded in the history of science and technology—is that von Neumann-style workshops have to preceded by a great many Babbage-style workshops. And there is no doubt that QIS has had a great many Babbage-style workshops … so is it the time now ripe for a von Neumann-style QIS workshop?
What the SMEC Committee had in-hand, that the QIS community does not presently possess (but soon will), is effective tools for system-level engineering. And therefore, as modern quantum system engineering tools come on-line, it is reasonable to foresee that these tools will catalyze an accelerating transition from Babbage-style QIS research to von Neumann-style QIS research.
And to conclude with a pragmatic question … isn’t a timely Babbage-to-von Neumann “phase transition” the only realistic hope that young researchers have of finding career employment in QIS? 🙂
Comment #34 April 9th, 2009 at 11:25 am
John, I find it ironic that you take Babbage as your prototype of a theorist–he was a great visionary, but NOT a theorist in the modern sense. Ada Lovelace might have a slightly stronger claim.
Comment #35 April 9th, 2009 at 11:45 am
Scott, it’s hard to argue that “modern theorist” has any single sense for QIS. After all, John von Neumann and Norbert Weiner were surely the founders of a modern school of theorists … but Alexander Grothendieck and Jean-Pierre Serre were founders of schools too, and their work definitely does have relevance to QIS.
IMHO, tolerance (and liking) for kind of cognitive diversity is absolutely essential to the health of QIS.
If you ask a QIT researcher to name a paradigmatic example of a “modern theorist” isn’t it true that a great many folks will point to a person who is very much like themselves?
Comment #36 April 9th, 2009 at 3:28 pm
Babbage-style fundamental QIS research — Does that mean engineering machines beyond the capability of current technology to build?
Babbage-style workshops — I’m at a loss to understand what this means. I thought Babbage mostly worked independently; and he worked without the benefit of any theoretical grounding, starting his computer designs 30 years before Boole’s An Investigation of the Laws of Thought.
Comment #37 April 9th, 2009 at 4:34 pm
Jack, doesn’t your first point provide a pretty reasonable answer to your second point? With the 2004 Quantum Computation Roadmap, Version 2.0 to flesh out the some of the details?
According to this 2004 Roadmap, the QIS community was projected to be on-track to a 2012 demonstration of “exercising multiple logical qubits through the full range of operations required for fault-tolerant QC” (which seems to me to be pretty much on-track) on “systems of the order of 50 physical qubits”.
Obviously, it is the latter milestone that is challenging from a technological point of view. We are a long way from the technological capability to create/control computational systems of 50 physically entangled qubits.
IMHO, there is one remarkable breakthrough that QIS can claim since 2004, which was not foreseen in the Roadmap. Namely, the growing appreciation that systems having state-space dimension or order 2^50 are simulatable with classical resources, provided the systems are symmetric, and/or low temperature, and/or noisy.
Because, from a quantum operations point of view, can’t these three simulatable quantum classes be viewed as being pretty much the same class?
E.g., isn’t a highly symmetric quantum state—like a ground-state ion in a Penning trap—simply a quantum state from which all the (informatically expensive) assymmetry has been extracted by an equivalent measurement process and/or contact with a thermal reservoir?
These fundamental QIS insights regarding simulatability—which have important near-term practical applications that span broad areas of materials science, quantum chemistry, molecular medicine, and electronics—are IMHO a major success of the QIS enterprise (as broadly conceived).
And yet, few in 2004 saw these insights coming … for example, the implications of QIS for practical ab initio quantum chemistry calculations are not foreseen in the 2004 Roadmap at all.
To the extend that the topic of the coming workshop is partly about whether QIS should be narrowly conceived (as focusing upon quantum computation), versus broadly conceived … well … that is the kind of topic upon which (IMHO) there should be a diversity of opinion.
Fortunately, it is reasonable to anticipate that invited speakers like Alán Aspuru-Guzik will do a good job of representing QIS as broadly conceived! 🙂
Comment #38 April 9th, 2009 at 7:51 pm
John says: “These fundamental QIS insights regarding simulatability—which have important near-term practical applications that span broad areas of materials science, quantum chemistry, molecular medicine, and electronics—are IMHO a major success of the QIS enterprise (as broadly conceived).”
Am I correct in thinking that a working Quantum Computer will be required for “practical applications”? Are there any QC’s that work outside of Canada? Can anyone provide an estimate of the rate of progress on achieving a working/useful QC compared to, say, the rate of progress on achieving controlled fusion in a working/useful tokamak?
Comment #39 April 9th, 2009 at 8:57 pm
Raoul Ohio asks: Am I correct in thinking that a working Quantum Computer will be required for “practical applications”?
It depends on whether you believe (as I do) that QIS has important practical applications—already—that do not require working quantum computers. And also believe (as I do) that QIS is intimately bound-up with wonderful opportunities in fundamental mathematics and physics that do not require that working quantum computers be constructable even in principle.
Isn’t it true, from this point of view, that the QIS community is in a win-win situation? Because a world in which large quantum systems (like biological molecules) are efficiently simulatable by QIS-inspired classical computation surely is just as wonderful to live in as a world in which large classical systems (like factoring engines) are simulatable by QIS-inspired quantum computation. And it may even be that we are all lucky enough live in both worlds. 🙂
Comment #40 April 9th, 2009 at 9:01 pm
John, good point!
Comment #41 April 9th, 2009 at 9:26 pm
John, thanks!
Comment #42 April 10th, 2009 at 3:13 am
Vienna is a lovely city and not too far; the list of participants and the topic look extremely interesting, and going there will be an homage of a sort to the late actor Groucho Marx. How very tempting.
Comment #43 April 10th, 2009 at 7:55 am
rrtucci,
I should clarify my earlier comment: I did not mean to somehow imply that there isn’t emerging commercial interest in simulation software for quantum/nonlinear optics, molecular dynamics, etc. In fact, I know folks who have formed startups in these areas. What I meant was that these people are almost entirely from academic backgrounds… or at least heavily involved with the academic community. Also, my impression is that this event is primarily geared towards building a strong theoretical community for QIS in the United States. This sort of long term investment/patronage is something the government can do better than private industry (I feel bad for saying this).
All of that aside, if there’s money to be made in short term software development (scale of a few years): (1) – it’ll happen without government support (which is ultimately poisonous), (2) – it’ll happen fast, (3) – people won’t want to talk about what they’re working on.
Comment #44 April 10th, 2009 at 9:39 am
KaoriBlue asserts: If there’s money to be made in short term software development (scale of a few years): (1) – it’ll happen without government support (which is ultimately poisonous), (2) – it’ll happen fast, (3) – people won’t want to talk about what they’re working on.
With respect, KaoriBlue, all three of those statements are either manifestly incorrect, or so weakly grounded in the history of science and technology as to be thoroughly misleading.
My experience at the 50th ENC was exactly the opposite: (1) essentially all quantum software development is occurring in the context of public-private hybrid organizations, (2) there is a pervasive feeling that relative to urgent global problems and expanding global opportunities, quantum software development is not happening fast enough, and (3) researchers are extremely happy to talk about what they were working on.
The most interesting quantum simulation software is all being released under GPL 2.0. The reason is simple. Simulation software—as broadly conceived—is no more naturally a commercial product than mathematics is a commercial product … rather, software (and mathematics) are best conceived, and put to their most productive use, as venues for team-building. And this creates a game in which open research and enterprise strategies are winning strategies.
Heck, why do you think that commercial electronics vendors so often release their compilers as free products? 🙂
Two very good books on this topic Burgelman and Grove (yes, that Grove), Strategy is Destiny, and Strategic Dynamics: Concepts and Cases. These works IMHO offer many useful lessons for the QIS community. The early history of IBM’s involvement with the space program is also highly instructive.
Comment #45 April 10th, 2009 at 10:52 am
Kaori Blue,
I too believe that private industry has a mind of its own and should only be weakly coupled to government funding and policy.
However, there is a coupling between academia and industry that is unavoidable, namely, that academia trains the scientists and engineers that propel industry. If there were no connection between academia and industry, then there would be, in my opinion, almost no justification for government to fund academic QIS research. I believe that government should try to enhance the ties between academia and industry. I don’t see how this can happen if industry is excluded or very under-represented from a workshop like this.
For example, why isn’t a business representative from Intel or Microsoft (a business type , not a pure mathematician like M. Freedman) speaking at this conference? Why wasn’t Geordie Rose invited to speak? Whatever you think about the prospects of D-Wave, it has garnered 40million dollars of investment, which is a pretty awesome achievement. It would also be nice to hear from investors and entrepreneurs who believe in QIS. They don’t have to reveal their trade secrets, just their opinions and advice. Why wasn’t John Sidles invited to speak (of course, his talk would have to have a very stringent time limit 🙂 ) Why wasn’t this advertised in Slashdot, to get the immensely powerful programming community involved? Why isn’t there some program, like a coop, to place QIS postdocs in industry? Why aren’t there tax incentives for companies working on QIS?
You say:
” my impression is that this event is primarily geared towards building a strong theoretical community for QIS in the United States.”
That’s my impression too. This workshop has been turned into just another theory conference. But, IMHO, this workshop should have and could have been something entirely different.
Comment #46 April 10th, 2009 at 11:17 am
John,
“Simulation software—as broadly conceived—is no more naturally a commercial product than mathematics is a commercial product … rather, software (and mathematics) are best conceived, and put to their most productive use, as venues for team-building. And this creates a game in which open research and enterprise strategies are winning strategies.”
First, I should say that I agree with most of what you’ve written, and I always appreciate a good rebuttal.
However, by “short term software development” I probably did a poor job describing what I meant. I’m talking more about people doing niche things like building faster/better/simpler modeling and simulation tools for a particular nanophotonic device they’re hoping to build, I’m talking about the group that finds a clever strategy for predicting the folding of particular types of hard-to-crystallize transmembrane or viral coat proteins with sufficient accuracy for rational drug design – structures they might want to sell (as a startup) to pharmaceutical outfits, a group (somewhere down the road) that wants to write software for faster/higher-resolution inverse lithography and sell it to chip designers, etc. Why did I mention these things? Because I think this is where VC’s, CEO’s, and non-academics kind of jump in – i.e. the people “not on the list” – and I’m not worried about these things happening without government support.
However, as I’ve mentioned, I do think that government (or hybrid government/large corporation) support is needed for the kind of comprehensive quantum simulation software development you’re talking about (as well as for more ambitious projects). As this proceeds, commercial offshoots will happen as they happen and eventually there will be a healthy industry.
Comment #47 April 10th, 2009 at 11:39 am
rrtucci,
“However, there is a coupling between academia and industry that is unavoidable, namely, that academia trains the scientists and engineers that propel industry. If there were no connection between academia and industry, then there would be, in my opinion, almost no justification for government to fund academic QIS research.”
I think there’s actually a pretty good coupling between academia and industry. My claim was actually that the people creating startups in QIS/quant. simulation/etc. are likely going to be from academia in the first place! For instance, at the institute I’m from, if you have a good idea it’s actually pretty easy to obtain a patent and launch a startup with funds from a variety of sources/grants/etc. and then interface with VC’s a little later on. As you know, patents can be a major source of revenue for universities/institutes, and there’s a clear mutualistic relationship if your startup succeeds.
Forgive me if I’m incorrect, but didn’t Geordie Rose start out this way?
“Why wasn’t this advertised in Slashdot, to get the immensely powerful programming community involved?”
To use a bit of hyperbole, isn’t that sort of like harnessing the collective brain power of middle school kids all over the nation to build a (compact) commercial fusion reactor?
“For example, why isn’t a business representative from Intel or Microsoft (a business type , not a pure mathematician like M. Freedman) speaking at this conference?”
I don’t think a regular suit from Microsoft would sit through the theorist’s talks and vice-versa. Besides, as I was saying above, I don’t think that these are the guys you should be concerned about when it comes to kick-starting an industry. When things mature a bit, maybe it would be good for Marburger to reach out to VC’s who specialize in QIS-like startups. There will be more as time goes on and they’ll make themselves known.
Comment #48 April 10th, 2009 at 12:19 pm
I remember when, in the late 1960s and early 1970s, conferences on Artificial Intelligence were lots of fun to attend. Very optimistic, interdisciplinary, one or two books from the past year being discussed, a handful of papers getting attention, encouraging “toy problems” being solved. Then there was that Winter, or multiple winters (the Lighthill Report in Europe, the Perceptrons book by Minsky and Papert) and in a few years, the conferences were filled with “suits” and premature commercialization and bullet charts with dollar signs. After a while, I stopped going. Then the Artificial Life and Nanotechnology conferences were lots of fun to attend. Very optimistic, interdisciplinary, one or two books from the past year being discussed, a handful of papers getting attention, encouraging “toy problems” being solved. Hence I share the concern of some commenters here that QIS is at risk of social pressures from its own (currently) exponential growth. Most fields follow something closer to a logistic curve. The optimum time for an IPO in the technology is right at the inflection point, where growth rate is greatest, and investors falsely extrapolate continued increasing growth. I am a QIS optimist, but it seems reasonable to have an analytic model that justifies an estimate of when that inflection point will occur.
Comment #49 April 10th, 2009 at 7:06 pm
Oh, so the suits caused the AI winter
Comment #50 April 10th, 2009 at 9:34 pm
Jonathan, forgive me for being flippant. Let me explain. The moral of your story seems to be that suits somehow caused the AI winters. I disagree with this. I believe AI winters are an inevitable part of science, and nobody is to blame for them. After all, scientific progress is a bunch of spikes. I fact, I believe suits can often help shorten the span of AI winters. Eg. the suits behind Celera probably reduced the time required to map the human genome by 10 years. And how about the suits behind RCA, Fairchild semiconductor, IBM, Intel, Bell Labs…
Comment #51 April 10th, 2009 at 9:55 pm
I should add that I too believe that suits can sometimes be deleterious to science. Eg. I’m a fan of Macia Angell, writer of “The Truth About the Drug Companies…”
Suits are just capitalism. As long as they are properly reigned in, suits can be highly beneficial to science and society.
Comment #52 April 10th, 2009 at 11:09 pm
rrtucci,
I’m kind of confused what you mean by ‘suits’? I was referring to people with MBA degrees and no real scientific training/background (a high level manager at Microsoft for example). I wouldn’t consider a VC to be a suit. And wasn’t Celera, along with many other biotech startups, founded by people with strong academic backgrounds?
Comment #53 April 11th, 2009 at 12:13 am
I was using the term suit, because Jonathan van Post used it. I guess what I meant by it is any member of a company that is trying to commercialize a scientific application.
Comment #54 April 11th, 2009 at 12:44 am
I enjoy following the evolution of the English language. Like many branches of science, it is much more interesting than a soap opera.
I think the term ‘suits’ originated in the music business. I understand it to refer to people who wear suits in hopes that others thinks they are a big deal.
Wearing a suit is not a totally negative thing. I encourage my students to buy a suit at Goodwill for a couple of bucks, and practice wearing it around the house, and while drinking a couple beers. That way they can go to a job interview or whatever, and not feel like an idiot because they are wearing a suit. If you are in a university and are considering moving from a professor to a dean or administrator, being able to look like you belong in a suit is a key element.
But, on the other hand, about 75% of the thrust of the word ‘suit’ is someone with a rudimentary knowledge about the issue at hand (music, CS, QIT, whatever) who thinks they should be regarded as a big deal only they are wearing a suit.
** tangent ** I have no idea about the suits as Microsoft, but I bet they are smarter than the average suit, because MS hires a lot of smart people. MS products usually get a bad rap in the CS world, but actually they are by far the best for “ordinary people”. For one thing, MS has usability labs where they find out what ordinary users can figure out. Everyone has heard plenty of bad raps about Vista, but if you actually try it out, in about 10 minutes you figure out the plan, and it is much better than the alternatives. System administration issues might be another story.
User interfaces are hard to make intuitive. I tell my students to make things work like the standard MS products. If you have a table, make it work like Excel. That way, anyone in the world can use it without thinking too much.
The extreme contrast to MS is GNU stuff, which is way powerful, but can only be used by a smart person who has spent a long time learning the “gnu way”. Unix/Linux is in a similar situation.
** /tangent **
Comment #55 April 11th, 2009 at 7:49 am
Stephen B. Johnson’s The Secret of Apollo: Systems Management in American and European Space Programs is a terrific book about “suits.” On a more personal level, so is Apollo flight engineer Gene Kranz’ book Failure is not an Option: Mission Control from Mercury to Apollo 13 and Beyond (both books can be previewed on Google).
Even from the Hollywood level, we have the fun, respectful, and historically accurate Apollo 13, which is highly recommended. Keep your eye out in the movie for the key role played by Apollo’s funky 70s-style system simulation technology … all of which was subcontracted to IBM! 🙂
In Johnson’s book we read a passage that should be thought-provoking for all the QIS Workshop attendees:
Does this sound familiar? To every QIS researcher, it should sound familiar. Because I think most people who contemplate the broad implications of QIS come to appreciate that it will play a central role in 21st century systems engineering … and in fact appreciate that at many companies, QIS (in its broad conception) is already playing a central role.
Where are the “Suits” who grasp and apply fundamental QIS principles? They are running (or soon will be running) every company whose products press against the fundamental limits to size, speed, sensitivity, and power efficiency.
Which is good news for young QIS researchers! 🙂
Researchers like Claude Shannon, John von Neumann, Norbert Weiner understand this confluence perfectly, and in fact all three are among the honored architects of modern system engineering (my recent short essay for PNAS documents their work’s connection to modern QIS objectives, via a systems engineering point of view).
Just to mention, two lines from Apollo 13 that today are resonant for everyone on our planet are EECOM Ted White’s sobering assessment “Gene, the Odyssey is dying” and Flight Controller Gene Kranz’ response, “What have we got on the spacecraft that’s good?” The strengthening capability of QIS is (IMHO) among the very best things that we have on the 21st century “spacecraft” that is our planet.
Systems engineers are famously optimistic, but not particularly humorous, and never are cynical. This reflects my attitude to QIS perfectly.
Comment #56 April 12th, 2009 at 5:04 am
(It is not the near-by Vienna… next time then. I hope Scott will update us regarding the new insights and answers emerging in Vienna, Virginia(!).)
Comment #57 April 12th, 2009 at 10:58 am
Since this thread is winding down, maybe now is a good time to mention that Ryan North’s Dinosaur Comics sells what no techno-nerd should be without: a Giclée-printed Time Traveler Essentials T-Shirt.
Let’s suppose, for example, that you are accidentally teleported back in time to the year 1900. You know that inventions like heavier-than-air flying machines are possible—which will let you control the flow of history—if only you can remember how these inventions work.
Not to worry! Your T-shirt has the answer … not only for airplanes, but for lightbulbs, penicillin, and even methods for determining the longitude. All in friendly, practical, nontechnical language … as the T-shirt puts it: “Energy is mass time the speed of light squared; take the credit.”
This suggests a strategy by which the Vienna QIS Workshop can enormously accelerate the progress of QIS: design and sell QIS T-shirts for time-travellers. Sooner or later, some time-travelling techo-nerd wearing one of these T-shirts will talk to the founders of QIS, who will instantly grasp and apply these ideas, thus accelerating the progress of QIS by decades.
What should the QIS Workshop put on its time-travel T-shirts? That’s a great topic for this forum! If we consult the Federal Vision for Quantum Information Science we find that it specifies three scoping questions.
Federal Scoping Q1: What is the true power of a general purpose quantum computer? This question is accompanied by a 1981 quote from Richard Feynman:
So we need to visit Feynman in 1981 wearing a T-shirt that says: “Richard, when it comes to open quantum systems, the arguments of your 1981 article simply don’t apply. In fact, the Connection Machine that you and your son are working on already has enough power to compute the properties of medium-sized molecules from ab initio quantum mechanics. This quantum simulation market will grow exponentially in coming decades, so that’s a good place to invest your resources.”
Federal Scoping Q2: Are there fundamental limits to our ability to control and manipulate quantum systems, and, if so, what constraints do they place on technology and QIS? This is accompanied by a 1959 quote from Richard Feynman:
This is a terrific opportunity for a T-shirt that says: “Richard, your 1959 article is wonderful, and you should push as hard as possible to turn all its ideas into hardware. Do this in three stages. First, start using magnetic resonance for imaging purposes. Even with 1959 technology you’ll be able to image living tissues with millimeter-scale resolution … from which you’ll create a whole new medical industry and make tens of billions of dollars. Invest these billions in very-large scale integrated circuits … because just imagine how many transistors could be printed on silicon at (say) 50 nanometer resolution, and how fast these transistors would switch. You’ll make hundreds of billions of dollars, and create a vibrant industry. Invest these hundreds of billions in quantum information science. Namely, combine quantum-limited magnetic resonance sensing with quantum-limited device fabrication, and commence observing (as both you and von Neumann envisioned) every structure in the biosphere with atomic resolution. This largest of all scientific projects will be an engine of prosperity and cooperation for the entire planet throughout much of the twenty-first century. Plus, you’ll make trillions of dollars, which might be fun. Don’t forget to work with Philip Anderson on this!”
Federal Scoping Q3: Are there new states of matter that emerge from collective quantum systems? This is accompanied by a 1972 quote from Philip Anderson:
In response, our time-travel T-shirt needs a paragraph that reads something like: “Philip, by 2009 there will be tens of thousands of articles published on the quantum theory of superconducting materials (many of these articles will be yours) yet there will still be no symmetry-driven consensus as to how these materials work … leading some folks to suspect that perhaps there never will be. The practical consequence will be that ab initio quantum simulation codes—embodying an increasingly deep understanding of quantum information science—will come to play an ever-increasing role in condensed matter physics, because these codes accommodate complex systems in which multiple competing physical mechanisms are in-play. Such systems will turn out to be ubiquitous in both materials science and biology. You should work with Richard Feynman on this!”
————————–
This leads to the interesting question, is the 2009 QIS Workshop chartered to suggest modifications of its own scoping questions? My tentative opinion is “yes” … for the pragmatic reason that the Federal Vision for Quantum Information Science was issued by the previous administration (and signed by political appointee John Marburger III). Nowadays we have a new OSTP director, John Holdren. So perhaps someone should ask Holdren’s office what the appropriate scoping questions are for QIS?
Comment #58 April 22nd, 2009 at 8:17 pm
Gosh, ten consecutive days with *zero* posts relating to the “Federal Vision for QIS”.
Perhaps this year-by-year survey of arxiv preprints whose abstract matches ‘(quantum AND (comput* OR information))’ will spark comment:
—————————-
Year: # arxiv preprints
2008: 739
2007: 734
2006: 738
2005: 712
2004: 683
2003: 645
2002: 534
2001: 475
2000: 390
1999: 278
1998: 215
1997: 147
—————————-
So, did the era of QIS growth really end circa 2004?
To my mind, QIS is assuming such central significance, to every branch of science and engineering—and mathematics too—that this stagnation (if it is real) is inexplicable/ regrettable/ strategically disastrous (your choice).