Monday, April 20, 2009

ACM-EC'09 papers

By now the list of accepted papers for ACM-EC is last week's news, already mentioned here and here. A nice feature of that list is that there is a version of it complete with abstracts, so we can get a better feel what what sort of papers appear there. I tried reading them; let me have a go at making comments (this year, I am an impartial observer, I did not submit a paper or serve on the program committee).

It's easy to classify papers according to who wrote them, harder to classify them according to (more importantly!) what they are about. Out of 40 papers, just over 31 come from the USA (dividing up credit for a paper equally amongst its co-authors), followed by Israel on just over 3, then England takes third place with just under 2, leaving France, Germany, etc simply nowhere.

There's a big emphasis on mechanism design, and most of the papers address the issue of truthful mechanisms in various contexts. Only one of the abstracts mentions price of anarchy; another paper is about price of mediation (ratio of social cost of correlated equilibria to social cost of Nash equilibria). Maybe that kind of topic is more likely to go to SAGT these days(?). (However, just to keep up the competitive pressure, (via the comsoc mailing list) I just got an announcement for AMMA, the first Conference on Auctions, Market Mechanisns and Their Applications.) Returning to EC, I see maybe 2 papers relating to coalitional games, another 2 to market equilibria, one on voting... actually, it's kind of hard to find a feature that is shared by approximately half of the papers. I wondered about classifying them according to the number of agents they study, i.e. 2, 3, n or infinity, but in nearly all cases it seems to be n.

1 comment:

Edith said...

How about applied/theoretical? Or experiments/proofs? To my taste, there are too many applied and experimental papers (these two are not necessarily the same), but maybe this is just sour grapes :)