Archive for August, 2010

AGT/E publication venues

Where is it best for an AGT/E researcher to publish his research?  How do we evaluate a CV in the AGT/E area?

As the number of young researchers in the AGT/E area increases, and as lower level hiring/promotion/tenure/grant committees need to “justify” their decisions to higher-up committees, it may be helpful to have some explicit discussion of the matter.  I am talking about the “domain-specific” knowledge of what carries a signal of excellence in AGT/E rather than the usual debates about the importance and pitfalls of various bibliometrics, rankings, recommendation letters, committees, etc, which obviously apply to AGT/E as well.

I would like to open this up for discussion, but let me start with my own impressions.

  1. AGT/E is interdisciplinary with most work currently coming from the following scientific communities (a) CS theory (b) AI (c) Networking (d) game theory (e) economic theory (f) operations research.  The best publishing venues can either belong to any one of these disciplines or can be in interdisciplinary forums.
  2. Judging disciplinary publishing venues is according to the standards in that field.  E.g. in economics the highest quality signal is publication in one of the “top 5″ journals (of which for AGT/E probably means in Econometrica).  Similarly, publications in top OR journals (or even general management science ones), or in the top GT journal, GEB, carry a signal of significance.  In the CS disciplines, conferences count most (FOCS/STOC and then SODA in CS theory, AAAI/IJCAI in AI, and SIGCOM and then INFOCOM in networking.)
  3. The top interdisciplinary forum in AGT/E is the ACM EC conference.  The top “web” conference WWWC may be considered another top interdisciplinary venue.  There is no serious AGT/E per-se  journal yet.  Other recognized well-regarded (but not top) truly AGT/E conferences are WINE and SAGT.
  4. In general, publication in a “top venue” in any one of the possible disciplines is a strong signal.  Still, not all these signals are of the same strength.  While it is hard to compare between the top forums in the different disciplines, at the risk (and hope) of drawing some anger among my readers, let me try my own (subjective) assessment of  the signal strength from the top venues (that I’m more or less familiar with) in the different types of disciplines:  Econometrica > STOC/FOCS > EC > WWWC > GEB > AAAI/IJCAI.
  5. Publishing  in a top conference and then the journal version in a top venue of another discipline is a plus.  An invited paper to a special journal issue collecting papers from a conference is a plus.
  6. Citation counts are more significant than the venue of publications.  Currently Google scholar is the citation count to use.


Read Full Post »

The organizers of the New York Computer Science and Economics Day asked me to post the following invitation for it:

We invite participants to the third New York Computer Science and Economics Day (NYCE Day), October 15 2010, at the New York Academy of Sciences, 7 World Trade Center. Details at: www.nyas.org/nyce2010

NYCE Day is a gathering for people in the larger New York City metropolitan area with interests in computer science, economics, marketing and finance to discuss common research problems and topics in a relaxed environment. The aim is to foster collaboration and the exchange of ideas.  The program features invited speakers Robert Almgren (NYU and Quantitative Brokers), Sham Kakade (Wharton), Robert Kleinberg (Cornell), Hal Varian (Google), and Mihalis Yannakakis (Columbia), and a rump session with short contributed presentations.

Call for Rump Session Speakers—Deadline September 13: We are currently soliciting speakers for the rump session. Each speaker will have 5 minutes to describe a problem and result, an experiment/system and results, or an open problem or a big challenge. If you would like to speak, please send a one paragraph text description of what you would like to present, including links to references if needed, to physicalscience@nyas.org with NYCE in the subject line. Submissions must be received by  Monday, September 13 and accepted speakers will be notified by September 20.

Your participation and suggestions are greatly welcome. Please pass this information on to people who may be interested.

NYCE Organizers
Christian Borgs, Microsoft

Michael Kearns, University of Pennsylvania
Sebastien Lahaie, Yahoo!
Vahab Mirrokni, Google

Read Full Post »

Swaprava Nath sent me a short summary that he wrote about the 11th Max Planck Advanced Courseon the Foundations of Computer Science on Approximation Algorithms and Algorithmic Game Theory that he recently attended in Saarbrücken, Germany:

The setting for this series of lectures were to discuss the computational issues involved in finding Nash Equilibrium and the hardness results. Given the hardness of the problems, approximation algorithms that are developed to address different relaxation of these problems were the focus of the talks. The speakers identified one area of these algorithmic issues.

The first one by Christos Papadimitriou addressed the general complexity classes and their interrelationships. The examples of the new complexity classes of PPAD, PPADS, PPA, PPP etc were shown with examples and showing that some of them implies some other were part of our exercises.

The second problem of Market Modeling and pricing of objects were addressed by Vijay Vazirani. In his lectures, the main focus was on Fisher’s linear market model and the Eisenberg Gale linear program which in the optimal solution gives a market clearing allocation and pricing (the allocation ensures that each bidder gets an optimal bundle of products and the pricing ensures that all products are sold and nobody has any surplus money).

The third problem of Cost sharing and approximation algorithms was addressed by Guido Schaefer. In his talks, he initially introduced the mechanism design framework and the reason for Incentive Compatibity. He discussed Moulin Mechanisms which under the assumption of ‘cross monotonic’ cost sharing functions can be shown to be group strategyproof and budget balanced. Note that all these frameworks work on a strong notion of Dominant Strategies. When more desirable properties of efficiency come into the play, because of the impossibility results of Green and Laffont, this calls for an approximation algorithm. The workaround in this setting is via approximating the budget balance and the total value. He analyzed these approximation algorithms with the example of Steiner Tree.

All the speakers concluded with some new directions in these fields and the current shape of the research. The lecture notes are available in the website.

Read Full Post »

Pricing Parking Spots in SF

The city of San Fransisco is experimenting with a new parking system that adjusts parking prices according to demand.  This demand is measured using newly installed sensors.  I predict a huge success, especially if they increase the upper bound of $6/hour on prices.

[Via NYT, Engadget, Cheaptalk]

Read Full Post »

During March 1 to August 31, 2011, the Institute of Advanced Studies at the Hebrew University of Jerusalem will host a group of leading researchers in Algorithmic Game Theory.

A workshop on “Innovations in Algorithmic Game Theory” will take place during May 22-27, 2011.

Read Full Post »

NSF ICES grant program

I recently got an email from NSF program director Tracy Kimbrel, addressed to several TCS bloggers, asking us to publicize the new NSF program on the Interface between Computer Science and Economics and Social Sciences.   While I have already posted about it, as well as about a report to NSF that may have partially motivated the program, I am flattered by being promoted from a blog writer to a blog editor, so here it is:

CISE has announced a new funding opportunity administered jointly with the Directorate for Social, Behavioral and Economic Sciences (SBE).  The new program is called ICES (Interface between Computer Science and Economics and Social Sciences).  You’ll find the all the details about ICES here:  http://www.nsf.gov/funding/pgm_summ.jsp?pims_id=503549.  Be sure to read the whole solicitation and click through to the non-exhaustive list of example topics as well.  ICES seeks to fund interdisciplinary work at one particular boundary between computer science and economics & social sciences.  As always with anything new at NSF, there will be lots of questions.   Email CISE/CCF Program Director Tracy Kimbrel (tkimbrel@nsf.gov), SBE/SES Program Director Nancy Lutz (nlutz@nsf.gov), CISE/IIS Program Director Sven Koenig (skoenig@nsf.gov), or CISE/CNS Program Director Darleen Fisher (dlfisher@nsf.gov) and we will do our best to answer the questions.  If you’re wondering “Is my project suitable for ICES?” be sure to include a one or two page description with your email because that will help us with the answer. The deadline for submitting proposals to ICES is October 5, 2010.

Read Full Post »