The goal of the project is to pick a handful of related papers representing a recent research trend in algorithmic mechanism design, read them, understand them, think really hard about future research directions in that area, and ideally embark on such research before the end of the semester. You may team up with a partner or two, depending on the size of the project you choose. I expect two things from you by the end of the semester
Take the next week to examine the list of suggestions below and decide on a project, then schedule a meeting with me by the end of next week (Friday 10/26) to discuss your choice and the composition of your team. Your team should ideally consist of two people, though it may be up to 3 people if you convince me that your project goals are particularly ambitious.
- Report: A report of at most 10 pages explaining what you learned, where you see this direction going, and any preliminary results or theorems you proved extending this work.
- Presentation: A 20-30 minute presentation to the class outlining the same.
Below you will find some ideas for projects, based on my (imperfect) sense of important current trends in algorithmic mechanism design. You can feel free to pick one of these after discussing with me (see above), or find your own (see below for suggestions on where to look).
NOTE: if any of the articles below are behind a paywall, simply append ".libproxy.usc.edu" to the domain portion of the URL (i.e. after .com, .org, etc)
to access it through the USC library.
- Revenue Optimal Auctions in Bayesian Multi-parameter Settings: Alaei, Alaei/Fu/Haghpanah/Hartline/Malekian (two papers), Cai/Daskalakis/Weinberg (two papers), Daskalakis/Weinberg
- Procurement auctions for private data: Ghosh/Roth, Roth/Ligett, Roth/Scheonebeck, Fleischer/Lyu
- Budgeted Procurement: Singer, Singer/Dobzinski/Papadimitriou, Chen/Gravin/Lu, Bei/Chen/Gravin/Lu
- Beyond Incentive Compatibility:
- Auctions in the presence of asymmetric information: Emek/Feldman/Gamzu/Leme/Tennenholtz, Alon/Feldman/Gamzu/Tennenholz, Babaioff/Kleinberg/Paes Leme
- Approximate Mechanism Design without Money: Procaccia/Tennenholz, Alon/Fischer/Procaccia/Tennenholz, Dughmi/Ghosh, Ashlagi/Fischer/Kash/Procaccia, and references therein.
- Other ideas: See the proceedings of EC 2012, EC 2011, EC 20??, and other conferences including STOC, FOCS, SODA, WINE, SAGT, ...