The Quantum Pontiff

“Ideal conversation must be an exchange of thought, and not, as many of those who worry most about their shortcomings believe, an eloquent exhibition of wit or oratory”
– Emily Post(er)

As a literature major physicist, one of the biggest culture shocks I’ve encountered when attending theory computer science conferences (STOC and FOCS) is the lack of a poster session at these conferences (or at least the ones I attended, which, truth be told, is not many.) Admittedly, I’m a sucker for free wine, beer, and cheese (or at least a cash bar peoples) and some of my warmest thoughts are of the science projects I did growing up (though I still think my eighth grade project was wrongly not awarded grand prize because the judges didn’t think I could have done the project.) But truthfully, I think I get more out of posters at conferences than most of the talks. And, in some deep sense, I find the lack of a poster session at these conferences nearly…anti-scientific. There. I’ve said it. Anti-scientific, Pontiff? Really? Well yeah I am prone to hyperbole.

Yes, I know that getting a paper into a top CS theory conference is the mark of acceptance and praise and “yes you are one of us” for the theoretical computer science community. And while I think that this system is inherently troubling for a few reasons, it doesn’t disturb me nearly as much as the suppression of ideas which are just “not good enough” or “too far on the fringe.” As far as I can tell, the inclusion of a poster session is strictly a positive: it gives students chance to discuss their work, it encourages breadth for a conference by allowing in submissions that might not fit with what is “in” at the moment, it is perhaps the best place I know to start a collaboration, and it encourages civil discussion of results that are…wrong.

Which brings me to my final point. I’ve been on the QIP program committee for two years now, and blessedly QIP does have a poster session (perhaps due to its hybrid nature combining physics and computer science.) This is great…for example in Santa Fe last year I learned about some very cool work the Dorit Aharonov and collaborators were doing from her poster (and a yelling match we had driving from ABQ to Santa Fe :)) as well as about some new results on the hidden subgroup problem over the Heisenberg group (okay I’ll admit that one is only really exciting to me!) In the course of these years we have, on the program committee, rejected posters from the conference. Now there are certain reason why I can imagine doing this, but it doesn’t really make me happy. For me, the only reason why one should reject a poster is that the poster should be off topic. For posters, because they’re just posters, damnit, I don’t even use the requirement that the paper is correct (sorry.) If QIP gets a paper on nonlinear fluid dynamics, then certainly reject the poster, but if you get a poster on P=NP? I say accept it as a poster (and I mean, it could be useful for the author to hear repeatedly why he or she is incorrect.)

So, if the QIP business meeting gets boring, and you want to stir up some debate, I suggest that someone raise the following question: what are our standards for poster session and are they ones that the conference should have? This might be especially relevant for QIP, which is increasingly computer science oriented (I said increasingly, not solely), and where there are certainly, say, implementation papers, that might get rejected (and I would argue they should be accepted: the beauty of quantum information science it’s crossing disciplines, and to cut of completely one discipline is like chopping your arm off.)

(One interesting issue for QIP is that for posters one submits the same 3 page brief note about the research that speakers submit. In most physics conferences when you submit a poster all you submit is an abstract, from which it is nearly impossible to judge whether the work is relevant/correct and so more posters are included.)

So, posters for all, says this scientific popularist.

Comments

  1. #1 John Sidles
    January 18, 2010

    Great post, Dave. Even supposing that a given poster is of interest to only (say) seven out of (say) 150 conference attendees — they might be the *right* seven. And one of the seven might be a student … who would be totally thrilled to engage in a long conversation … *and* have a beer!

    Seriously, I don’t think I went to a single conference in 2009 that *didn’t* have a poster session. Heck, the 50th ENC was more than 95% posters — and it was a truly *great* conference. Why is CS anomalous?

  2. #2 Steve Flammia
    January 18, 2010

    The poster session this year at QIP has some really excellent work. At one point today, I actually found myself saying, “I can’t believe how many of these posters I want to see!” And what is good about a poster is that you can query the author interactively, making results much easier to verify and understand.

  3. #3 Ian Durham
    January 18, 2010

    One quick correction: I’m pretty sure for QIP, posters only required one page as opposed to the three required for talks. Also, APS policy is that no one is rejected, even for oral talks (anyone who has been to a quantum foundations session from the early days of GQI remembers several “interesting” talks…).

  4. #4 Dave Bacon
    January 18, 2010

    Thanks Ian for the correction.

    The APS conference are fine, as far as things go, but I don’t find them a good model: 10-15 minute talks are nearly useless. I like the model of posters for all, talks by those that a committee has deemed interesting.

  5. #5 Matthew Putman
    January 19, 2010

    This really does make sense. It is like science blogs really. Blogs provide a contrast to peer review. Both are important.

  6. #6 Eric Lund
    January 19, 2010

    10-15 minute talks are nearly useless

    Scientifically speaking, ITA. As a practical matter, there comes a point where one needs to sit down for a while, and going to a talk or three is a perfect excuse for doing so.

    My big meeting tends to be AGU, which nominally has an acceptance policy but in practice anyone who complies with the one-contributed-paper-per-first-author policy (with a second allowed in education or public policy, or if invited) is accepted (and it’s easy to game the policy by permuting the author list). But the meeting is typically about 2/3 posters and 1/3 talks. I spend most of the meeting on the poster floor. But standing or walking on the concrete floor of the Moscone Center gets to the joints after a few hours, so it’s nice to pick a few talks to strategically attend and rest the bones.

  7. #7 Jeff Watson
    January 23, 2010

    Posters are the best thing about doing science projects. I’ll bet my folks saved the posters from every science project I ever did. I also remember not winning a prize in 6th grade, because they said “No way I could have done it.”

    Jeff