Friday, January 08, 2010

COLT and CCC

The COLT (Computational Learning Theory) call for papers is out. (Actually its been out since October but I was only recently emailed it.) For other information about COLT see here.
How do COLT and CCC relate to each other? (I use COLT for both the conference and the field. I use CCC for both the conference and the field.)
  1. There are some results in COLT that are of interest to CCC and vice versa. But there is not much overlap. That is, the papers at COLT would be out-of-scope at CCC. And vice versa.
  2. CCC has its original roots in computability theory. The basic notions of reductions and completeness were adapted from computability theory. COLT has some roots in Inductive Inference (computability-theoretic model of learning) but the connection is much weaker. The PAC model, and the other models, do not really take things from Inductive Inference and adapt them.
  3. Both conferences used to have more of the Computability-theoretic material but it is faded in recent years.
  4. Both fields use tools from discrete math; however, virtually all of Theoretical Computer Science uses discrete math.
  5. COLT is co-located with ML (Machine Learning). They have done this before (I'm not sure how often.) COLT would like to be relevant to ML and probably is. CCC does not have a (more) applied field that it would like to be relevant to. CCC co-locating with STOC since there is a overlap in the people who want to go to both.
COLT is in Israel. Is this a good idea? A conference should be located in a place where the following occur.
  1. There are people who normally can't go but now CAN go (e.g., CCC in Prague has 12(?) people from Prague, most of whom normally would have a hard time going). So- are there people in Israel who want to go? I would think yes since there is a strong theory community.
  2. It is not too hard for the people who usually go to go. How hard will it be for Americans to get to Israel? For Europeans? I don't know.
  3. The Guest Speakers (in this case Noga Alon and Noam Nissan) are close by thus saving on travel expenses. Actually, I had never thought of this one until I saw that they were the guest speakers.

6 comments:

  1. GARSARCH: "CCC does not have a (more) applied field that it would like to be relevant to."

    CCC CHARTER: "The CCC encourages results from other areas of computer science and mathematics motivated by computational complexity theory."

    ??? ... a serious question ... is this an aspect of the culture of complexity theory that can only be parsed by insiders?

    ReplyDelete
  2. How hard will it be for Americans to get to Israel? For Europeans?

    Don't forget Iranians.

    ReplyDelete
  3. Random comment: Has anyone been blogging from ICS the last couple days?

    ReplyDelete
  4. Sidles: YES, if some other area is doing work that is motivated by complexity,
    CCC would consider it in scope. Perhaps some problem in CCC leads to work in combinatorics or COLT,
    then work in combinatorics or COLT would be in scope.

    Of course, this is hypothetical. I'm do not know
    any examples off hand, though
    other people might.

    ANON: I mentioned Americans and Europeans.
    I should have said
    ``how hard will it be for the people who usually go to the conference to go this time
    now that its in Israel''
    ALSO- I hope people in that part of the world who usually can't go, can go this time.

    OTHER ANON: We have some guest blogs from ICS lined up.

    ReplyDelete
  5. COLT is co-located with ML (Machine Learning). They have done this before (I'm not sure how often.)

    COLT is colocated with ICML this year; many years it is also colocated with ICML and UAI.

    ReplyDelete
  6. In theory, CCC could have applications in cryptography.

    ReplyDelete