Tutorials

EC’22 will host virtual tutorials from June 28th to July 1st (the same week as the contributed poster sessions, which will take place between the two parts of the tutorials). You must register for the conferences in order to access the virtual venue. A link to the venue can be found here.

Tutorials will be presented live in the virtual venue. Each tutorial will have two parts. Part 1 will be presented 10:30-11:15am ET, and part 2 will be presented 12:00-12:45pm ET. The recording of the tutorials will be replayed 8:30-9:15pm ET and 10:00-10:45pm ET. The tutorial co-chairs are Shengwu Li and Nisarg Shah.

The tutorials are:

June 28: Redistributive Market Design. Website.
Speakers: Mohammad Akbarpour, Piotr Dworczak, and Scott Duke Kominers

Many scarce public resources are allocated at below-market-clearing prices—and sometimes for free. Such “non-market” mechanisms necessarily sacrifice some surplus, yet they can potentially improve equity. In this tutorial, we show how tools that have been developed in the past 40 years in mechanism design theory can be employed to identify the optimal structure of such redistributive policies. This work sheds light on how and when it may make sense to use non-market allocation systems, such as rationing and priority mechanisms.
June 29: Total Search Problems in Economics and Computation.
Speakers: Aris Filos-Ratsikas, Paul W. Goldberg, and Alexandros Hollender

Total search problems (i.e., problems for which a solution is guaranteed to exist) are ubiquitous in economics and computation, with several prominent examples coming from game theory, competitive markets and fair division. In this tutorial we will highlight the major results of the related literature on these problems and we will present the main techniques for analyzing their computational complexity. The aim of the tutorial is to introduce these concepts to the wider EC audience and to enable researchers to engage in research related to these problems.
June 30: Economics of Distributed Systems.
Speakers: Jacob Leshno and Matt Weinberg

Distributed systems allow a collection of many computers to act as one system. Doing so requires consensus between the different computers within the system. A rich literature in computer science addresses this problem from both theoretical and practical perspectives. These algorithms allow simultaneous editing of a Google Doc by different users, streaming of Netflix movies from multiple datacenters, and the reliability of the Bitcoin ledger of payments.

The goal of this tutorial is to provide an introduction and an overview of the theory behind consensus protocols, including an overview of modeling approaches, impossibility results, and practical and simple algorithms. The tutorial will cover classic results in distributed systems (FLP impossibility, Paxos) as well as more recent developments related to decentralized systems and cryptocurrencies (for example, the proof-of-work protocol that underlies Bitcoin). Recent developments raise questions about the feasibility of consensus among independent selfish entities. In particular, the design of consensus protocols is intertwined with mechanism design questions. Answering these questions requires a combination of tools from both economics and computer science, and the tutorial will provide such an introduction.
July 1: Learning-Augmented Mechanism Design.
Speakers: Eric Balkanski and Vasilis Gkatzelis

This tutorial will introduce the model of “learning-augmented mechanism design” (or “mechanism design with predictions”), which is an alternative model for the design and analysis of mechanisms in strategic settings. Aiming to complement the traditional approach in computer science, which ana- lyzes the performance of algorithms based on worst-case instances, recent work on “algorithms with predictions” has developed algorithms that are enhanced with machine-learned predictions regarding the optimal solution. The algorithms can use this information to guide their decisions and the goal is to achieve much stronger performance guarantees when these predictions are accurate (consistency), while also maintaining good worst-case guarantees, even if these predictions are very inaccurate (robustness).

So far, most of these results have been limited to online algorithms but some very recent work has shown that a possibly even more fertile ground for this model is in mechanism design. This tutorial will cover the foundations of learning-augmented mechanism design, some recent results in this model, and the many exciting directions for future-work in this area. In particular, we will focus on employing the learning-augmented mechanism design framework in the areas of 1) mechanism design without money, 2) the design of decentralized cost-sharing mechanisms, 3) auction design, and 4) online mechanism design.