Monday, August 30, 2010

New Institute for Theory of Computing

The Simons Foundation has announced a competition to establish a new Institute for the Theory of Computing in the United States.
Computation (and its abstract form, the algorithm) has not only revolutionized science, technology, and society, but also is among the most important scientific concepts discovered and developed in the 20th century. This scientific discipline has enabled numerous technological advances and has forged many connections to mathematics and other sciences, providing fruitful insights and new problems. It has impacted not only computer science and technology, but also parts of mathematics, physics, biology, economics and sociology. Meanwhile, its core scientific agenda is extremely ambitious and challenging. In short, this theoretical field is one of the most exciting and important today, attracting excellent young talent to its ranks at a growing rate. Young people with education and training in this field are well positioned to make central contributions to computer science and science in general.
An institute focused on the theory of computation could bring together a critical mass of researchers from around the world to accelerate fundamental research on computation and to further develop its interactions with other areas of science ranging from mathematics and statistics to biology, physics and engineering. The Simons Foundation invites applications for grants to establish such an Institute.
Letter of intent due by October 27 and full proposals by next June.

It's the funding that catches the eye, $6 million/year for 10 years with possible renewed funding or endowment after that. To understand the scale, that's roughly the budget of TTI-Chicago. This new institute will play a major role in our field.

The institute will be modeled on the Kavli Institute for Theoretical Physics, MSRI and the IMA (also similar to DIMACS) which are located at or near major university campuses and host programs on a given topic for several months to a year with long-term visitors and have several workshops related to those program topics.

This is opposed to the Oberwolfach/Dagstuhl/BIRS model of an isolated location with on-site room and board, weekly workshops but little to no long term programs or visitors. I'm glad the Simons Foundation is going with the former model. While I've enjoyed the many Dagstuhl workshops I've attended, centers like DIMACS tend have a greater long term impact on the field.

Where will this institute be located? We'll find out next fall. Should be exciting.

13 comments:

  1. I believe this is fundamental for the progress of the theory of computing. I am not sure, but I believe there is not an institute that stands up as a leader in Computational Theory. I believe we need such an institute, not in order to dominate the field, but in order to combine the efforts of scientist around the world. Worldwide, theory of computing should be embedded in CS courses in a way similar to mathematics, while IT studies should be a smaller part of CS degrees, or different degrees completely, as I believe it is being done in the USA.

    ReplyDelete
  2. Does the field really need many more postdoc positions, without any additional more-permanent research positions?

    ReplyDelete
  3. More money in the field is always a good thing. What are natural locations for such an institute?

    Princeton already has IAS
    MIT already has MSR

    Pittsburgh? Berkeley? Palo Alto? New York?

    ReplyDelete
  4. I wonder: is there any chance that the institute will not be hosted at [or near] one of MIT, Berkeley, CMU, Princeton, or Stanford?

    The rich get richer...

    ReplyDelete
  5. Anon above: It makes sense. If you are hosting postdocs, they want to be near serious concentrations of top theory people. The places you list (probably just MIT, CMU, Princeton, Berkeley actually) are the only such concentrations of serious researchers. Sure, other universities might have 1 or 2 each, but not enough to support an institute.

    ReplyDelete
  6. If you are hosting postdocs, they want to be near serious concentrations of top theory people.

    This is true, but there is also an argument for establishing a new concentration of theory people as opposed to making an existing one stronger.

    ReplyDelete
  7. That's fantastic! KITP is an extraordinary institution in the physics world, and establishing a similar place in the theory of computation would be huge...

    ReplyDelete
  8. While I've enjoyed the many Dagstuhl workshops I've attended, centers like DIMACS tend have a greater long term impact on the field.

    Any data to back this up, or is this something subjective?

    ReplyDelete
  9. "Princeton has IAS."

    IAS is not an institute for theoretical computer science. It is primarily a math institute, and wrt to TCS, more specifically a complexity institute.

    And CCI is just temporary. So Princeton would be a great place to have such an institute.

    ReplyDelete
  10. This is true, but there is also an argument for establishing a new concentration of theory people as opposed to making an existing one stronger.

    Did the TTI, with a comparable budget, have a large impact or a small impact on the field?

    If the former, then a new location would be called for, if the latter, then it might be best to double up the money of some pre-existing facility (TTI, DIMACS, IAS-CS-branch), etc.

    ReplyDelete
  11. MIT does not have MSR. They are completely separate although close
    geographically.

    Also MSR intended mission is to do game theory and economics and not core theory.

    ReplyDelete
  12. But look at who MSR has hired, Adam, Yael, Madhu, Boaz.. to an outside observer it seems that their intended mission is to do complexity theory :)

    ReplyDelete