Michal Feldman Reports from WINE 09

Michal Feldman wrote  a short report from WINE 09:

The workshop on Internet and network economics was held last week in the Sapienza University of Rome. The conference was extremely well organized by PC chair Stefano Leonardi, and the papers’ quality was high relative to previous years. The research areas were very diverse, more or less a representative sample of research in algorithmic game theory, including, among others, mechanism design, congestion games, strong equilibria, prediction markets, fair allocations and ad auctions. It seems that there is no longer a real distinction between WINE and SAGT in topics, participants and quality.

Peyton Young gave a keynote talk on adaptive learning in systems of interactive agents, which surveyed a family of adaptive learning rules, in which a Nash equilibrium will be player a large proportion of the time as an unintended consequence of individual adaption.

The session on envy-free mechanisms included papers both in the traditional cake-cutting (divisible) setting and the setting of indivisible goods with payments. David Kempe introduced an interesting model of envy-free mechanisms under budget constraints, and showed positive algorithmic results.

Angelina Vidali presented her work on the geometry of truthfulness, where she introduces a new geometrical method for characterizing the allocation graph of truthful mechanisms. This work may prove useful in improving the lower bound of the job scheduling problem with unrelated machines.

Nicole Immorlica gave a talk on externalities in keyword auctions, where she presented empirical results, based on Microsoft Live, indicating that externality effects (i.e., the ads in the other sponsored positions) are statistically significant.

Advertisement