Friday, August 31, 2007

ARC funding for theory projects

ARC (Australian Research Council) Discovery projects are the main source of funding for theoretical computer science research in australia. ARC publishes all data about the projects that are funded including the level of funding per year. It is not too easy to figure out exactly which projects should be considered as "theory" but one way would be to use the RFCD codes that ARC uses to classify research areas.

ARC funds about 4 Discovery projects per year in the RFCD code 2804 that seems to be the closest to theoretical computer science. Projects run for 3 years. In the last 2 years the total budget awarded was about $1m to these theory projects. That's about $85K per project per year. Each project has an average of 2.6 names on it. Which means that the average funding is about $32K per project per year per person named on the proposal. It could be that about 30 theorists have an ARC discovery project running. Since it is allowed to have a maximum of 2 discovery projects at the same time, there should be between 15 and 30 theorists with an ARC discovery running at the moment.

The Computer science research area is within the RFCD code 280000-Information, computing and communication sciences. The "theory" RFCD code should be 2804-Computation theory and mathematics. The 280000 ICT area has 5 RFCD codes/sub-areas as shown in the table below. Those are the areas that actually received funding in the last 3 years.

In the past 3 years, 2005-7 (2007 means "funding awarded in 2006 for projects beginning in 2007") the ARC funded proposals as follows:

codeArea200720062005
2801 Information systems182316
2802AI and signal processing2416
20
2803software5510
2804comp theory4410
2805Data format587
2800All ICT- total (280000)565664*


Source: ARC website. * in 2005 there was one more proposal funded in the "Other ICT" category, RFCD code 2899.

In 2007 the 4 projects in the theory RFCD code were the following:
  1. Efficient Pre-Processing of Hard Problems: New Approaches, Basic Theory and Applications, Griffith University, Prof V Estivill-Castro; Dr MR Fellows; Prof MA Langston, $380K in 4 years
  2. A Virtual Electromagnetic Compatibility (EMC) Lab Based on Advanced Computer Modeling and Simulation Techniques, Griffith University, A/Prof J Lu; Dr E Li, $275K in 3 years
  3. Information security and digital watermarking with Latin squares, Monash University, Dr TE Hall; Dr IM Wanless; Dr AZ Tirkel, $223K in 3 years
  4. Application of novel exact combinatorial optimisation techniques and metaheuristic methods for problems in cancer research, The University of Newcastle, Dr PA Moscato; Prof RJ Scott; Dr MA Langston, $237K in 3 years

In 2006 ("funding begins in 2006") the 4 projects in the theory RFCD code were the following:
  1. Early detection of component incompatibility in time-dependent computer architectures, University of Adelaide, DG Hemer, $157K in 3 years
  2. Fast, practical and effective algorithms for clustering with advice, University of Melbourne, Dr AI Wirth, $153K in 3 years
  3. A Grid-Enabled Meta-Server for Protein Threading, University of Sydney, Prof AY Zomaya; Dr B Zhou; Dr M Charleston; Dr A Viglas, $366K in 3 years
  4. Foundations of Nonmonotonic Logic Programming for Complex Knowledge Systems, University of Western Syndey, A/Prof Y Zhang; Dr AC Nayak; Dr K Wang; A/Prof F Lin, $234K in 3 years
In 2005 ("funding begins in 2005") the 10 projects in the theory RFCD code appear in the list below. Note that 4 of these projects may not be directly related to computer science theory.
  1. An innovative computational technique for the study and control of oscillation marks in continuous casting of steel, Curtin University of Technology, A/Prof Y Wu, $190K in 3 years
  2. Coarse Grained Parallel Algorithms, Griffith University, Prof FK Dehne, $206K in 3 years
  3. Algebraic reasoning for serialisability in probabilistic transaction systems, Macquarie University, Dr AK McIver; Dr CC Morgan; Prof T Nipkow, $147K in 3 years
  4. Devising sophisticated computational comparative genomic analysis strategies for animal and plant genomes, Murdoch University, Prof M Bellgard; A/Prof JK Kulski; Prof R Appels; Dr RH Taplin; A/Prof RR Barrero, $160K in 3 years
  5. THE DEVELOPMENT OF MECHANISTIC MODELS FOR BUBBLY FLOWS WITH HEAT AND MASS TRANSFER, RMIT University, A/Prof JY Tu; Dr GH Yeoh; Prof G Park, $178K in 3 years
  6. Advanced computational techniques for micro/nano multiscale systems of NEMS/BioMEMS, The University of Sydney, Dr Y Gu, $222K in 3 years
  7. Optimum design of controlled drug delivery systems, The University of Western Australia, A/Prof S Wang; Dr X Lou, $231K in 3 years
  8. Knowledge Based Model Updating for the Correctness of Security Protocols, University of Western Sydney, A/Prof Y Zhang; A/Prof MA Orgun; Dr AC Nayak; Dr Y Mu; Dr F Bao, $284K in 3 years
  9. Timeless digital signature for self-organising groups, University of Wollongong, Prof R Safavi-Naini; Prof PR Wild, $278K in 3 years
  10. Short Signatures: Tools for Securing Digital Transactions, and Their Applications, University of Wollongong, Dr W Susilo; Dr Y Mu; Dr F Zhang, $192K in 2 years
RFCD codes

the RFCD (Research Fields, Courses and Disciplines Classification) codes for computer science are organized as follows:
  • 2800 INFORMATION, COMPUTING AND COMMUNICATION SCIENCES
    1. 280100 Information Systems
    2. 280200 Artificial Intelligence and Signal and Image Processing
    3. 280300 Computer Software
    4. 280400 Computation Theory and Mathematics
      1. 280401 Analysis of Algorithms and Complexity
      2. 280402 Mathematical Logic and Formal Languages
      3. 280403 Logics and Meanings of Programs
      4. 280404 Numerical Analysis
      5. 280405 Discrete Mathematics
      6. 280406 Mathematical Software
      7. 280499 Computation Theory and Mathematics not elsewhere classified
    5. 280500 Data Format
    6. 289900 Other Information, Computing and Communication Sciences

Thursday, August 30, 2007

on CATS

I thought that the registration fee for CATS2007 was totally ridiculous. Complaining about conference fees is very common indeed, but still...
here is a comparison of registration fees in some of the best theory conferences around:

registration fees (late registration):
FOCS 2007: 435 USD = 528 AUD
STOC 2007: 425+100(FCRC fee)=525USD = 637 AUD (516AUD when it's not in the FCRC)
SoCG 2007: 283 EUR = 468 AUD
CCC'06: 300 EUR = 496 AUD

CATS07: 900 AUD

What could be the explanation for this?
IWOCA seems to be more reasonable with a registration at 380AUD. CATS is organized as part of the ACSW multi-conference event and is overseen by Computer Research and Education Association (CORE - previously CSA).

Perhaps IWOCA should be the "local" theory conference of choice?

Thursday, August 23, 2007

Australian CS Theory Conferences

There are two annual CS theory conferences in Australia - CATS and IWOCA (formerly known as AWOCA). Both seem to struggle with their identity. Should they aim to become internationally renowned conferences or should they mainly focus on bringing researchers together from the region? I know there are drawbacks with both choices, but imho they should focus on the latter.

First of all there are already too many conferences of questionable standards that claim to be prestigious, secondly I don't see anything wrong with being a regional conference whose aim is to bring Asia-Pacific researchers together. I always try to attend CATS since I think it's important for the Australian CS theory community to be active, however, I'm not keen on submitting a paper that could be accepted to a much better conference, especially not if any of my co-authors is a student.

In Europe the computational geometry community have had big success organising the European Workshop on CG (EWCG). The format allows everyone to submit a four page abstract that is reviewed for correctness. Almost all the papers are accepted but they are not published in any official proceedings, thus allowing the authors to submit their paper to more prestigious conferences. The success of the workshop shows that this is a possible (even recommendable?) format. The majority of the European CG community usually attend so it's a great place to meet all colleagues.

I heard people say that this format isn't possible because they can't get funding for it. However, I don't see this is a problem. Especially not since Australian academia seems to be all about networking and very little about doing good research.