Archive for the ‘Complexity’ Category

Quantum Complexity Theory Student Project Showcase 2!

Saturday, December 22nd, 2012

(Note: The “2!” in the title of this post really does mean “2 factorial,” if you want it to.)

With the end of the semester upon us, it’s time for a once-every-two-year tradition: showcasing student projects from my 6.845 Quantum Complexity Theory graduate course at MIT.  For my previous showcase, in 2010, I chose six projects that I thought were especially outstanding.  This year, however, there were so many great projects—and so many, in particular, that could actually be useful to people in quantum computing—that I decided simply to open up the showcase to the whole class.  I had 17 takers; their project reports and 10-minute presentation slides are below.

Let me mention a few projects that tried to do something new and audacious.  Jenny Barry generalizes the notion of Partially Observable Markov Decision Processes (POMDPs) to the quantum case, and uses a recent result of Eisert et al., showing that certain problems in quantum measurement theory are undecidable (like, literally Turing-undecidable), to show that goal state reachability for “QOMDPs” is also Turing-undecidable (despite being decidable for classical POMDPs).  Matt Falk suggests a novel quantum algorithm for spatial search on the 2D grid, and gives some numerical evidence that the algorithm finds a marked item in O(√n) time (which, if true, would be the optimal bound, beating the previous runtime of O(√(n log n))).  Matt Coudron and Henry Yuen set out to prove that the Vazirani-Vidick protocol for quantum randomness expansion is optimal, and achieve some interesting partial results.  Mohammad Bavarian (well, jointly with me) asks whether there’s a fast quantum algorithm for PARITY that gets the right answer on just slightly more than 50% of the inputs—and shows, rather surprisingly, that this question is closely related to some of the hardest open problems about Boolean functions, like sensitivity versus block sensitivity.

This year, though, I also want to call special attention to the survey projects, since some of them resulted in review articles that could be of real use to students and researchers in quantum computing theory.  Notably, Adam Bookatz compiled the first list of essentially all known QMA-complete problems, analogous to (but shorter than!) Garey and Johnson’s listing of known NP-complete problems in 1979.  Chris Graves surveyed the known quantum fault-tolerance bounds.  Finally, three projects took on the task of understanding and explaining some of the most important recent results in quantum complexity theory: Travis Hance on Thomas Vidick and Tsuyoshi Ito’s NEXP in MIP* breakthrough; Emily Stark on Mark Zhandry’s phenomenal results on the security of classical cryptographic constructions against quantum attack; and Max Zimet on Jordan-Lee-Preskill’s major work on simulation of quantum field theories.

(Oops, sorry … did I use words like “important,” “breakthrough,” and “phenomenal” too often in that last sentence, thereby triggering the wrath of the theoretical computer science excitement police?  Well then, come over to my apartment and friggin’ arrest me.)

Anyway, thanks so much to all the students for making 6.845 such an awesome class (at least on my end)!  Without further ado, here’s the complete project showcase:

• Jenny Barry.  Quantum POMDPs (Partially Observable Markov Decision Processes).  [Report] [Slides]
• Matt Coudron and Henry Yuen.  Some Limits on Non-Local Randomness Expansion.  [Report] [Slides]
• Badih Ghazi.  Quantum Query Complexity of PARITY with Small Bias.  [Report] [Slides]
• Chris Graves.  Survey on Bounds on the Quantum Fault-Tolerance Threshold.  [Report] [Slides]
• Travis Hance.  Multiprover Interactive Protocols with Quantum Entanglement.  [Report] [Slides]
• Vincent Liew.  On the Complexity of Manipulating Quantum Boolean Circuits.  [Report] [Slides]

The Boson Apocalypse

Friday, December 21st, 2012

If the world ends today, at least it won’t do so without three identical photons having been used to sample from a probability distribution defined in terms of the permanents of 3×3 matrices, thereby demonstrating the Aaronson-Arkhipov BosonSampling protocol.  And the results were obtained by no fewer than four independent experimental groups, some of whom have now published in Science.  One of the groups is based in Brisbane, Australia, one in Oxford, one in Vienna, and one in Rome; they coordinated to release their results the same day.  That’s right, the number of papers (4) that these groups managed to choreograph to appear simultaneously actually exceeds the number of photons that they so managed (3).  The Brisbane group was even generous enough to ask me to coauthor: I haven’t been within 10,000 miles of their lab, but I did try to make myself useful to them as a “complexity theory consultant.”

Here are links to the four experimental BosonSampling papers released in the past week:

For those who want to know the theoretical background to this work:

For those just tuning in, here are some popular-level articles about BosonSampling:

I’ll be happy to answer further questions in the comments; for now, here’s a brief FAQ:

Q: Why do you need photons in particular for these experiments?

A: What we need is identical bosons, whose transition amplitudes are given by the permanents of matrices.  If it were practical to do this experiment with Higgs bosons, they would work too!  But photons are more readily available.

Q: But a BosonSampling device isn’t really a “computer,” is it?

A: It depends what you mean by “computer”!  If you mean a physical system that you load input into, let evolve according to the laws of physics, then measure to get an answer to a well-defined mathematical problem, then sure, it’s a computer!   The only question is whether it’s a useful computer.  We don’t believe it can be used as a universal quantum computer—or even, for that matter, as a universal classical computer.  More than that, Alex and I weren’t able to show that solving the BosonSampling problem has any practical use for anything.  However, we did manage to amass evidence that, despite being useless, the BosonSampling problem is also hard (at least for a classical computer).  And for us, the hardness of classical simulation was the entire point.

Q: So, these experiments reported in Science this week  have done something that no classical computer could feasibly simulate?

A: No, a classical computer can handle the simulation of 3 photons without much—or really, any—difficulty.  This is only a first step: before this, the analogous experiment (called the Hong-Ou-Mandel dip) had only ever been performed with 2 photons, for which there’s not even any difference in complexity between the permanent and the determinant (i.e., between bosons and fermions).  However, if you could scale this experiment up to about 30 photons, then it’s likely that the experiment would be solving the BosonSampling problem faster than any existing classical computer (though the latter could eventually solve the problem as well).  And if you could scale it up to 100 photons, then you might never even know if your experiment was working correctly, because a classical computer would need such an astronomical amount of time to check the results.

Proving Without Explaining, and Verifying Without Understanding

Saturday, November 17th, 2012

Last Friday, I was at a “Symposium on the Nature of Proof” at UPenn, to give a popular talk about theoretical computer scientists’ expansions of the notion of mathematical proof (to encompass things like probabilistic, interactive, zero-knowledge, and quantum proofs).  This really is some of the easiest, best, and most fun material in all of CS theory to popularize.  Here are iTunes videos of my talk and the three others in the symposium: I’m video #2, logician Solomon Feferman is #3, attorney David Rudovsky is #4, and mathematician Dennis DeTurck is #5.  Also, here are my PowerPoint slides.  Thanks very much to Scott Weinstein at Penn for organizing the symposium.

In other news, the Complexity Zoo went down yet again this week, in a disaster that left vulnerable communities without access to vital resources like nondeterminism and multi-prover interaction.  Luckily, computational power has since been restored: with help from some volunteers, I managed to get the Zoo up and running again on my BlueHost account.  But while the content is there, it looks horrendously ugly; all the formatting seems to be gone.  And the day I agreed to let the Zoo be ported to MediaWiki was the day I lost the ability to fix such problems.  What I really need, going forward, is for someone else simply to take charge of maintaining the Zoo: it’s become painfully apparent both that it needs to be done and that I lack the requisite IT skills.  If you want to take a crack at it, here’s an XML dump of the Zoo from a few months ago (I don’t think it’s really changed since then).  You don’t even need to ask my permission: just get something running, and if it looks good, I’ll anoint you the next Zookeeper and redirect complexityzoo.com to point to your URL.

Update (Nov. 18): The Zoo is back up with the old formatting and graphics!!  Thanks so much to Charles Fu for setting up the new complexity-zoo.net (as well as Ethan, who set up a slower site that tided us over).  I’ve redirected complexityzoo.com to point to complexity-zoo.net, though it might take some time for your browser cache to clear.

The Toaster-Enhanced Turing Machine

Thursday, August 30th, 2012

Over at Theoretical Computer Science StackExchange, an entertaining debate has erupted about the meaning and validity of the Church-Turing Thesis.  The prompt for this debate was a question asking for opinions about Peter Wegner and Dina Goldin’s repetitive diatribes claiming to refute “the myth of the Church-Turing Thesis”—on the grounds that, you see, Turing machines can only handle computations with static inputs and outputs, not interactivity, or programs like operating systems that run continuously.  For a demolition of this simple misunderstanding, see Lance Fortnow’s CACM article.  Anyway, I wrote my own parodic response to the question, which generated so many comments that the moderators started shooing people away.  So I decided to repost my answer on my blog.  That way, after you’re done upvoting my answer over at CS Theory StackExchange :-), you can come back here and continue the discussion in the comments section.

Here’s my favorite analogy. Suppose I spent a decade publishing books and papers arguing that, contrary to theoretical computer science’s dogma, the Church-Turing Thesis fails to capture all of computation, because Turing machines can’t toast bread. Therefore, you need my revolutionary new model, the Toaster-Enhanced Turing Machine (TETM), which allows bread as a possible input and includes toasting it as a primitive operation.

You might say: sure, I have a “point”, but it’s a totally uninteresting one. No one ever claimed that a Turing machine could handle every possible interaction with the external world, without first hooking it up to suitable peripherals. If you want a Turing machine to toast bread, you need to connect it to a toaster; then the TM can easily handle the toaster’s internal logic (unless this particular toaster requires solving the halting problem or something like that to determine how brown the bread should be!). In exactly the same way, if you want a TM to handle interactive communication, then you need to hook it up to suitable communication devices, as Neel discussed in his answer. In neither case are we saying anything that wouldn’t have been obvious to Turing himself.

So, I’d say the reason why there’s been no “followup” to Wegner and Goldin’s diatribes is that theoretical computer science has known how to model interactivity whenever needed, and has happily done so, since the very beginning of the field.

Update (8/30): A related point is as follows. Does it ever give the critics pause that, here inside the Elite Church-Turing Ivory Tower (the ECTIT), the major research themes for the past two decades have included interactive proofs, multiparty cryptographic protocols, codes for interactive communication, asynchronous protocols for routing, consensus, rumor-spreading, leader-election, etc., and the price of anarchy in economic networks? If putting Turing’s notion of computation at the center of the field makes it so hard to discuss interaction, how is it that so few of us have noticed?

Another Update: To the people who keep banging the drum about higher-level formalisms being vastly more intuitive than TMs, and no one thinking in terms of TMs as a practical matter, let me ask an extremely simple question. What is it that lets all those high-level languages existin the first place, that ensures they can always be compiled down to machine code? Could it be … err … THE CHURCH-TURING THESIS, the very same one you’ve been ragging on? To clarify, the Church-Turing Thesis is not the claim that “TURING MACHINEZ RULE!!” Rather, it’s the claim that any reasonable programming language will be equivalent in expressive power to Turing machines — and as a consequence, that you might as well think in terms of the higher-level languages if it’s more convenient to do so. This, of course, was a radical new insight 60-75 years ago.

Update (Sept. 6): Check out this awesome comment by Lou Scheffer, describing his own tale of conversion from a Church-Turing skeptic to believer, and making an extremely apt comparison to the experience of conversion to the belief that R, R2, and so on all have the same cardinality (an experience I also underwent!).

Complexity Zoo is down — anyone willing to help?

Tuesday, July 31st, 2012

Update (August 5): Sorry for the delay! Now that the Zoo is back up, my sense of urgency has decreased, but we still do need a long-term solution. Thanks so much to everyone who offered hosting. Alas, I was persuaded by the argument that it’s too complicated to have a wiki mirrored at multiple locations, so I should really choose one—and ideally it should be someplace where I retain control of the files, in case anything goes wrong again. Following the helpful directions of Eric Price, I set up a MediaWiki installation at http://scottaar.scripts.mit.edu/zoo. Is anyone interested in helping me transfer over the content from the qwiki Zoo?

Update (August 1): Thanks to the efforts of Gopal Sarma at Stanford, the Zoo is back up and running!!  However, I believe the only long-term solution is to get the Zoo mirrored at other locations.  I can then direct the domain complexityzoo.com to point to any of them that are currently up.  So, to all of those who volunteered to mirror the Zoo: thanks so much, and please go ahead and do so!  Let me know what you need for that (I can ask Gopal to get the source files).

As some of you have noticed, the Complexity Zoo (well, don’t bother clicking the link!) has been down for the past couple weeks.  Some Stanford students volunteered to host the Zoo years ago but then graduated, and these sorts of outages have been a frustrating reality since then.  So my co-zookeeper Greg Kuperberg and I are looking for a volunteer to help us get the Zoo back online.  The reward?  Eternal gratitude and a co-zookeeper title for yourself.  In principle, I could host the Zoo on my Bluehost account, but I don’t know how to set up the wiki software, and I’m not even sure how to retrieve the Zoo pages prior to its going down (Google Cache?).  If you’re interested or have ideas, leave a comment or send me an email.

Thanks!!

Big news

Thursday, March 15th, 2012

Judea Pearl has won a richly-deserved Turing Award, for his pioneering work on reasoning under uncertainty, Bayesian inference, and causality.  Much like last year’s winner Leslie Valiant, Pearl has been a perfectly-plausible candidate since the 1980s; it was really just a question of when they’d get around to him.  For those who don’t know his work, Pearl’s landmark book Causality provides a wonderful introduction to at least one major strand of his thought; I read it this summer and it inverted the way I think about lots of things in statistics.  (Pearl’s fame precedes this award, partly for a tragic reason: he’s probably best known to the public as the father of the murdered journalist Daniel Pearl.)

In other big news, playing Super Mario Bros. is now known to be NP-complete, as shown in this landmark paper by Greg Aloupis, Erik Demaine, and Alan Guo.  The sheer intuitiveness of the gadget constructions, at least to anyone who grew up playing Nintendo, makes this probably my favorite NP-completeness paper of all time (well, I guess tied with some papers by Cook, Karp, and Levin).

My visit to D-Wave: Beyond the roast-beef sandwich

Tuesday, February 21st, 2012

Last week I was in Vancouver, to give talks at the University of British Columbia and at the American Association for the Advancement of Science annual meeting.  As part of that visit, on Friday afternoon, John Preskill, John Martinis, Michael Freedman and I accepted a gracious invitation to tour the headquarters of D-Wave Systems in Burnaby (a suburb of Vancouver).  We started out in a conference room, where they served us cookies and sodas.  Being the mature person that I am, the possibility of the cookies being poisoned at no point crossed my mind.

Then we started the tour of D-Wave’s labs.  We looked under a microscope at the superconducting chips; we saw the cooling systems used to get the chips down to 20 millikelvin.  In an experience that harked back to the mainframe era, we actually walked inside the giant black cubes that D-Wave was preparing for shipment.  (The machines are so large partly because of the need for cooling, and partly to let engineers go in and fix things.)  Afterwards, D-Wave CTO Geordie Rose gave a 2-hour presentation about their latest experimental results.  Then we all went out to dinner.  The D-Wave folks were extremely cordial to us and fielded all of our questions.

In spite of my announcement almost a year ago that I was retiring as Chief D-Wave Skeptic, I thought it would be fitting to give Shtetl-Optimized readers an update on what I learned from this visit.  I’ll start with three factual points before moving on to larger issues.

Point #1: D-Wave now has a 128-(qu)bit machine that can output approximate solutions to a particular NP-hard minimization problem—namely, the problem of minimizing the energy of 90-100 Ising spins with pairwise interactions along a certain fixed graph (the “input” to the machine being the tunable interaction strengths).  So I hereby retire my notorious comment from 2007, about the 16-bit machine that D-Wave used for its Sudoku demonstration being no more computationally-useful than a roast-beef sandwich.  D-Wave does have something today that’s more computationally-useful than a roast-beef sandwich; the question is “merely” whether it’s ever more useful than your laptop.  Geordie presented graphs that showed D-Wave’s quantum annealer solving its Ising spin problem “faster” than classical simulated annealing and tabu search (where “faster” means ignoring the time for cooling the annealer down, which seemed fair to me).  Unfortunately, the data didn’t go up to large input sizes, while the data that did go up to large input sizes only compared against complete classical algorithms rather than heuristic ones.  (Of course, all this is leaving aside the large blowups that would likely be incurred in practice, from reducing practical optimization problems to D-Wave’s fixed Ising spin problem.)  In summary, while the observed speedup is certainly interesting, it remains unclear exactly what to make of it, and especially, whether or not quantum coherence is playing a role.

Which brings me to Point #2.  It remains true, as I’ve reiterated here for years, that we have no direct evidence that quantum coherence is playing a role in the observed speedup, or indeed that entanglement between qubits is ever present in the system.  (Note that, if there’s no entanglement, then it becomes extremely implausible that quantum coherence could be playing a role in a speedup.  For while separable-mixed-state quantum computers are not yet known to be efficiently simulable classically, we certainly don’t have any examples where they give a speedup.)  Last year, as reported on this blog, D-Wave had a nice Nature paper that reported quantum tunneling behavior in an 8-qubit system.  However, when I asked D-Wave scientist Mohammad Amin, he said he didn’t think that experiment provided any evidence for entanglement between qubits.

The “obvious” way to demonstrate entanglement between qubits would be to show a Bell inequality violation.  (We know that this can be done in superconducting qubits, as the Schoelkopf group at Yale among others reported it a couple years ago.)  Meanwhile, the “obvious” way to demonstrate a role for quantum coherence in the apparent speedup would be gradually to “turn down” the system’s coherence (for example, by adding an interaction that constantly measured the qubits in the computational basis), and check that the annealer’s performance degraded to that of classical simulated annealing.  Unfortunately, the D-Wave folks told us that neither experiment seems feasible with their current setup, basically because they don’t have arbitrary local unitary transformations and measurements available.  They said they want to try to demonstrate 2-qubit entanglement, but in the meantime, are open to other ideas for how to demonstrate a quantum role in the apparent speedup with their existing setup.

Point #3: D-Wave was finally able to clarify a conceptual point that had been bugging me for years.  I—and apparently many others!—thought D-Wave was claiming that their qubits decohere almost immediately (so that, in particular, entanglement would almost certainly never be present during the computation), but that the lack of entanglement didn’t matter, for some complicated reason having to do with energy gaps.  I was far from alone in regarding such a claim as incredible: as mentioned earlier, there’s no evidence that a quantum computer without entanglement can solve any problem asymptotically faster than a classical computer.  However, that isn’t D-Wave’s claim.  What they think is that their system decoheres almost immediately in the energy eigenbasis, but that it doesn’t decohere in the computational basis—so that, in particular, there would be entanglement at intermediate stages.  If so, that would be perfectly fine from the standpoint of the adiabatic algorithm, which doesn’t need coherence in the energy eigenbasis anyway (after all, the whole point is that, throughout the computation, you want to stay as close to the system’s ground state as possible!).  I understand that, given their knowledge of decoherence mechanisms, some physicists are extremely skeptical that you could have rapid decoherence in the energy basis without getting decoherence in the computational basis also.  So certainly the burden is on D-Wave to demonstrate that they maintain coherence “where it counts.”  But at least I now understand what they’re claiming, and how it would be compatible (if true) with a quantum speedup.

Let me now move on to three broader questions raised by the above points.

The first is: rather than constantly adding more qubits and issuing more hard-to-evaluate announcements, while leaving the scientific characterization of its devices in a state of limbo, why doesn’t D-Wave just focus all its efforts on demonstrating entanglement, or otherwise getting stronger evidence for a quantum role in the apparent speedup?  When I put this question to Mohammad Amin, he said that, if D-Wave had followed my suggestion, it would have published some interesting research papers and then gone out of business—since the fundraising pressure is always for more qubits and more dramatic announcements, not for clearer understanding of its systems.  So, let me try to get a message out to the pointy-haired bosses of the world: a single qubit that you understand is better than a thousand qubits that you don’t.  There’s a reason why academic quantum computing groups focus on pushing down decoherence and demonstrating entanglement in 2, 3, or 4 qubits: because that way, at least you know that the qubits are qubits!  Once you’ve shown that the foundation is solid, then you try to scale up.  So, please support D-Wave if it wants to spend money to show Bell inequality violations, or other “smoking-gun” evidence that its qubits are working together coherently.  You’re welcome, D-Wave!

The second question is one that I’ve encountered many times on the blogosphere: who cares how D-Wave’s system works, and whether it does or doesn’t exploit quantum coherence, as long as it solves practical problems faster?  Sure, maybe what D-Wave is building is really a series of interesting, useful, but still basically “classical” annealing devices.  Maybe the word “quantum” is functioning here as the stone in a stone soup: attracting money, interest, and talented people to build something that, while neat, ultimately doesn’t much depend on quantum mechanics at all.  As long as D-Wave’s (literal!) black box solves the problem instances in such-and-such amount of time, why does it matter what’s inside?

To see the obtuseness of this question, consider a simple thought experiment: suppose D-Wave were marketing a classical, special-purpose, $10-million computer designed to perform simulated annealing, for 90-bit Ising spin glass problems with a certain fixed topology, somewhat better than an off-the-shelf computing cluster. Would there be even 5% of the public interest that there is now? I think D-Wave itself would be the first to admit the answer is no. Indeed, Geordie Rose spoke explicitly in his presentation about the compelling nature of (as he put it) “the quantum computing story,” and how it was key to attracting investment. People don’t care about this stuff because they want to find the ground states of Ising spin systems a bit faster; they care because they want to know whether or not the human race has finally achieved a new form of computing. So characterizing the device matters, goddammit! I pride myself on being willing to adjust my opinions on just about anything in response to new data (as I’ve certainly done in D-Wave’s case), but the insistence that black boxes must be opened and explanations provided is something I’ll carry to the grave. Finally, given the skeptical-yet-positive tone of this post, some people will wonder whether I now regret my earlier, more unmitigated D-Wave skepticism. The answer is no! Asking questions is my job. I’ll give D-Wave credit whenever it answers some of the questions—as it did on this visit!—and will shift my views accordingly. But I’ll also neither stop asking nor apologize for asking, until the evidence for a quantum speedup becomes clear and indisputable (as it certainly hasn’t yet). On the other hand, I do regret the snowballing nastiness that developed as a combined result of my and other skeptics’ statements, D-Wave’s and its supporters’ statements, and the adversarial nature of the blogosphere. For the first time, I find myself really, genuinely hoping—with all my heart—that D-Wave will succeed in proving that it can do some (not necessarily universal) form of scalable quantum computation. For, if nothing else, such a success would prove to the world that my$100,000 is safe, and decisively refute the QC skeptics who, right now, are getting even further under my skin than the uncritical D-Wave boosters ever did.

Whether or not God plays dice, I do

Friday, February 3rd, 2012

Another Update (Feb. 7): I have a new piece up at IEEE Spectrum, explaining why I made this bet.  Thanks to Rachel Courtland for soliciting the piece and for her suggestions improving it.

Update: My $100,000 offer for disproving scalable quantum computing has been Slashdotted. Reading through the comments was amusing as always. The top comment suggested that winning my prize was trivial: “Just point a gun at his head and ask him ‘Convinced?’” (For the record: no, I wouldn’t be, even as I handed over my money. And if you want to be a street thug, why limit yourself to victims who happen to have made public bets about quantum computing?) Many people assumed I was a QC skeptic, and was offering the prize because I hoped to spur research aimed at disproving QC. (Which is actually an interesting misreading: I wonder how much “pro-paranormal” research has been spurred by James Randi’s million-dollar prize?) Other people said the bet was irrelevant since D-Wave has already built scalable QCs. (Oh, how I wish I could put the D-Wave boosters and the QC deniers in the same room, and let them duke it out with each other while leaving me alone for a while!) One person argued that it would be easy to prove the impossibility of scalable QCs, just like it would’ve been easy to prove the impossibility of scalable classical computers in 1946: the only problem is that both proofs would then be invalidated by advances in technology. (I think he understands the word “proof” differently than I do.) Then, buried deep in the comments, with a score of 2 out of 5, was one person who understood precisely: I think he’s saying that while a general quantum computer might be a very long way off, the underlying theory that allows such a thing to exist is on very solid ground (which is why he’s putting up the money). Of course this prize might still cost him since if the news of the prize goes viral he’s going to spend the next decade getting spammed by kooks. OK, two people: There’s some needed context. Aaronson himself works on quantum complexity theory. Much of his work deals with quantum computers (at a conceptual level–what is and isn’t possible). Yet there are some people who reject the idea the quantum computers can scale to “useful” sizes–including some very smart people like Leonid Levin (of Cook-Levin Theorem fame)–and some of them send him email, questions, comments on his blog, etc. saying so. These people are essentially asserting that Aaronson’s career is rooted in things that can’t exist. Thus, Aaronson essentially said “prove it.” It’s true that proving such a statement would be very difficult … But the context is that Aaronson gets mail and questions all the time from people who simply assert that scalable QC is impossible, and he’s challenging them to be more formal about it. He also mentions, in fairness, that if he does have to pay out, he’d consider it an honor, because it would be a great scientific advance. For better or worse, I’m now offering a US$100,000 award for a demonstration, convincing to me, that scalable quantum computing is impossible in the physical world.  This award has no time limit other than my death, and is entirely at my discretion (though if you want to convince me, a good approach would be to convince most of the physics community first).  I might, also at my discretion, decide to split the award among several people or groups, or give a smaller award for a discovery that dramatically weakens the possibility of scalable QC while still leaving it open.  I don’t promise to read every claimed refutation of QC that’s emailed to me.  Indeed, you needn’t even bother to send me your refutation directly: just convince most of the physics community, and believe me, I’ll hear about it!  The prize amount will not be adjusted for inflation.

The impetus for this prize was a post on Dick Lipton’s blog, entitled “Perpetual Motion of the 21st Century?”  (See also this followup post.)  The post consists of a debate between well-known quantum-computing skeptic Gil Kalai and well-known quantum-computing researcher Aram Harrow (Shtetl-Optimized commenters both), about the assumptions behind the Quantum Fault-Tolerance Theorem.  So far, the debate covers well-trodden ground, but I understand that it will continue for a while longer.  Anyway, in the comments section of the post, I pointed out that a refutation of scalable QC would require, not merely poking this or that hole in the Fault-Tolerance Theorem, but the construction of a dramatically-new, classically-efficiently-simulable picture of physical reality: something I don’t expect but would welcome as the scientific thrill of my life.  Gil more-or-less dared me to put a large cash prize behind my words—as I’m now, apparently, known for doing!—and I accepted his dare.

To clarify: no, I don’t expect ever to have to pay the prize, but that’s not, by itself, a sufficient reason for offering it.  After all, I also don’t expect Newt to win the Republican primary, but I’m not ready to put $100,000 on the line for that belief. The real reason to offer this prize is that, if I did have to pay, at least doing so would be an honor: for I’d then (presumably) simply be adding a little to the well-deserved Nobel Prize coffers of one of the greatest revolutionaries in the history of physics. Over on Lipton’s blog, my offer was criticized for being “like offering$100,000 to anyone who can prove that Bigfoot doesn’t exist.”  To me, though, that completely misses the point.  As I wrote there, whether Bigfoot exists is a question about the contingent history of evolution on Earth.  By contrast, whether scalable quantum computing is possible is a question about the laws of physics.  It’s perfectly conceivable that future developments in physics would conflict with scalable quantum computing, in the same way that relativity conflicts with faster-than-light communication, and the Second Law of Thermodynamics conflicts with perpetuum mobiles.  It’s for such a development in physics that I’m offering this prize.

Update: If anyone wants to offer a counterpart prize for a demonstration that scalable quantum computing is possible, I’ll be happy for that—as I’m sure, will many experimental QC groups around the world.  I’m certainly not offering such a prize.

Cerebrum-stuffer from Shtetl Claus

Saturday, December 24th, 2011

Ho3!  Home with family for the holidays and looking for something to do?  Then check out the archives of our 6.893 Philosophy and Theoretical Computer Science course blog.  The course just ended last week, so you can find discussions of everything from the interpretation of quantum mechanics to Occam’s Razor to the Church-Turing Thesis to strong AI, as well as links to student projects, including Criticisms of the Turing Test and Why You Should Ignore (Most of) Them, Barwise Inverse Relation Principle, Bayesian Surprise, Boosting, and Other Things that Begin with the Letter B, and an interactive demonstration of interactive proofs.  Thanks to my TA Andy Drucker, and especially to the students, for making this such an interesting course.

My New York Times essay on quantum computing

Monday, December 5th, 2011

I have a special treat for those commenters who consider me an incorrigible publicity-hound: an essay I was invited to write for the New York Times Science section, entitled Quantum Computing Promises New Insights, Not Just Supermachines.  (My original title was “The Real Reasons to Study Quantum Computing.”)  This piece is part of a collection of essays on “the future of computing,” which include one on self-driving cars by Sebastian Thrun, one on online learning by Daphne Koller, and other interesting stuff (the full list is here).

In writing my essay, the basic constraints were:

(a) I’d been given a rare opportunity to challenge at least ten popular misconceptions about quantum computing, and would kick myself for years if I didn’t hit all of them,

(b) I couldn’t presuppose the reader had heard of quantum computing, and