Talks
Fall 2015

Decision Making at Scale: Platforms and Problems

Wednesday, November 18th, 2015 10:40 am11:20 am

Add to Calendar

While the Internet has dramatically changed many socio-economic interactions, there are still no widely used platforms for large-scale substantive discussion on complex societal issues, or for large scale participatory democracy. The lack of such platforms makes it hard to formulate algorithmic problems and design mechanisms which would, in turn, accelerate the development and deployment of these platforms. In this talk, we will describe some of our preliminary attempts at breaking this cycle by describing two systems that we have built and deployed for participatory democracy, and outlining some of the algorithmic and mechanism design questions that arise.
 
We will first present a voting platform for participatory budgeting. Most budgeting problems are knapsack problems, and hence most participatory budgeting problems can be reduced to knapsack voting. We will describe partial progress towards incentive-compatible and pragmatic voting rules for knapsack voting. We will also outline some research directions in the area of knapsack voting in particular, and structured decision making (where the votes and the decisions admit a simple formal description) in general.
 
We will then describe an experiment in unstructured decision making, where a characterization of the decision space is not known a priori. We will use this to outline general design principles for these systems, to highlight the need for "consensus markets" (akin to "prediction markets"), and present a candidate design that holds promise but doesn't quite work.