Wednesday, July 04, 2007

FOCS 07 list out

The list is here. As I was instructed by my source, let the annual "roasting of the PC members" begin !

63 papers were accepted, and on a first look there appears to be a nice mix of topics: it doesn't seem as if any one area stands out. Not many papers are online though, from my cursory random sample, so any informed commenting on the papers will have to wait. People who know more about any of the papers are free to comment (even if you're the author!). Does anyone know the number of submissions this year ? I heard it was quite high.

19 comments:

  1. My rejection notification says 66 accepted papers among 302 submissions. :-D

    ReplyDelete
  2. Gosh, an acceptance rate of 22%! This must have made for some really tough choices. I was given three preprints by various authors containing strong results and all three got rejected. Good thing the SODA deadline is not yet past us.

    ReplyDelete
  3. 63 papers were accepted...

    My rejection notification says 66 accepted papers...

    Some papers got merged.

    ReplyDelete
  4. For example, having one data structures paper in the conference is a nice mix of topics :)

    ReplyDelete
  5. From the titles it looks like about 1/3rd of the papers are algorithmic. This seems contrary to the "algorithmic lense" perspective put forth by Papadimitriou as part of the CCC effort to increase funding for CS.

    ReplyDelete
  6. From the titles it looks like about 1/3rd of the papers are algorithmic. This seems contrary to the "algorithmic lense" perspective put forth by Papadimitriou as part of the CCC effort to increase funding for CS.

    I'm not sure I understand. I guess you mean that 1/3 is "too low" ? Ignoring usual arguments about what is "algorithmic", it's fair to say that we're not going to see changes immediately, and secondly the measure of success of the 'algorithmic lens' is certainly not going to be the number of "algorithmic papers" at FOCS (or STOC) but a broader enriching of areas outside theory,(and even mention of things like NP-hardness etc in areas far removed from the CS community)

    ReplyDelete
  7. the measure of success of the 'algorithmic lens' is certainly not going to be the number of "algorithmic papers" at FOCS (or STOC)

    In other words "do as I say not as I do".

    (and even mention of things like NP-hardness etc in areas far removed from the CS community)

    Is that the best we have to offer to the outside community: "sorry, can't do. it's NP complete".

    I sure hope not!

    Our best contributions from practice are when we go and say "how long does it take you to do X?... well we can do it for you and order of magnitude faster!". This applies equally for problems in P and NP complete ones (presumably not in P). Suppose the naive solution takes time O(n^a 2^(n^b)) but we can reduce the a and b in the expression through clever algorithms. Say if we start with a=4, b=2 and we have a cpu limit time of 1,000,000 the largest n that is effective would be around 2. For a=4, b=1/2 we could solve problems for n up to 16 and for a=1, b=1/2 we could handle n=160.

    ReplyDelete
  8. The committee is commendable for how quickly they got reviews out to authors--most pleasant reviews for a rejected paper I've ever gotten. (Plus that perennial favorite, "I didn't read this carefully, but it sucks".)

    ReplyDelete
  9. In response to the "In other words "do as I say not as I do".
    " anon,

    that's misreading my comment rather dramatically, as well as misunderstanding the point of FOCS/STOC: they (for better or for worse) are supposed to attract the most groundbreaking results in whatever areas are deemed that year to be core theory.

    An algorithm that is used to improve the running time of a scheme and save 1000s of seconds is great, and a spotlight should be focused on such activities, but that doesn't mean that FOCS/STOC are the right places for such work. The Kanellakis prize honors theory work that has a substantial nontrivial impact on an area of practice, and much of the awarded research has its theory side and a practical side, the latter not necessarily showing up in *any* publication.

    ReplyDelete
  10. There are no learning theory papers either.
    Nice mix of topics indeed.

    ReplyDelete
  11. To Mihai and last anonymous:

    I suppose that the mix should depend on what is submitted in each area. Do you guys know of good papers (say more than a couple) in data structures or learning theory that were rejected?

    ReplyDelete
  12. Do you guys know of good papers (say more than a couple) in data structures or learning theory that were rejected?

    For all we know no learning papers were submitted but this doesn't imply that the mix of topics is thus good. It simply means that this suboptimal mix happened through no fault of the PC.

    ReplyDelete
  13. On a different topic: SODA08 server seems to have died shortly before the 5pm deadline (or at least it does not like my emails ;) Anyone else having this experience ?

    ReplyDelete
  14. Why accept 66 papers?

    It is impossible to argue that accepting another ten would harm overall quality or make the printed proceedings too big.

    This just postpones the scientific dissemination of very strong work for another 6 or 12 months. Who benefits from this?

    (I did not have a submission in this FOCS, so the above is not sour grapes.)

    ReplyDelete
  15. It's a simple calculation: a full day of talks, single session, for three days. you can't go above 63 talks unless you're willing to go to multiple sessions, or sacrifice invited talks. FOCS has historically been single track.

    So the correct question really is: does FOCS serve the community better as a single track or multiple track conference. Maybe you should bring it up at the business meeting.

    ReplyDelete
  16. It's a simple calculation: a full day of talks, single session, for three days. you can't go above 63 talks unless you're willing to go to multiple sessions, or sacrifice invited talks.

    The accepts at FOCS'98 and FOCS'02 broke 75 (see http://3dpancakes.typepad.com/ernie/2005/06/stocky_lies_and.html). I think FOCS'02 was single session.

    ReplyDelete
  17. Oops. Shows how much I know. FOCS has had parallel sessions for a while now. THe main difference I see between the two cited FOCSs (98,02) and now (06, say) is that those two had three "tutorials" and no invited talks (and it is possible that the tutorials were on a different day). I should remember 98, since I was there, but I can't seem to recollect whether the tutorials were before/after the conference or during.

    In any case, that might explain it, together with the special session occasionally used for award papers.

    ReplyDelete
  18. I think that FOCS 2002 had *parallel* sessions, see this link

    ReplyDelete
  19. Is FOCS2007.org still an active webpage? The link doesn't work.

    ReplyDelete

Disqus for The Geomblog