https://www.dagstuhl.de/19381

September 15 – 20 , 2019, Dagstuhl Seminar 19381

Application-Oriented Computational Social Choice

Organizers

Umberto Grandi (University Toulouse Capitole, FR)
Stefan Napel (Universität Bayreuth, DE)
Rolf Niedermeier (TU Berlin, DE)
Kristen Brent Venable (IHMC – Pensacola, US)

For support, please contact

Dagstuhl Service Team

Documents

Dagstuhl Report, Volume 9, Issue 9 Dagstuhl Report
Aims & Scope
List of Participants
Shared Documents
Dagstuhl's Impact: Documents available
Dagstuhl Seminar Schedule [pdf]

Summary

Computational social choice (COMSOC) combines models from political science and economics with techniques from computer science, to analyze collective decision processes from a computational perspective. Classical contributions include the study of the computational barriers to various forms of manipulation in elections, the definition of novel procedures for distributed resources among a group of human or artificial agents, as well as the study of complex collective decisions such as multi-winner voting rules and voting in combinatorial domains. COMSOC is a thriving field of research, with an international bi-annual workshop now at its 7th edition and a handbook published in 2016 which structures more than a decade of research, but future success will depend on the practical applicability of its findings. The purpose of this seminar was to address this challenge by stimulating application-driven research in computational social choice, i.e., theoretical studies modeling existing practical problems in all their complexity.

Four areas of COMSOC, which have already proven or bear particular potential for synergies and applicability to real-life problems, were identified as the focus of the seminar. Each of these areas addresses present-day challenges that provide an opportunity for an interdisciplinary approach building on contributions from computer scientists, economists, mathematicians, and political scientists:

  • Recommender systems is a very successful application that combines several artificial intelligence techniques. Indeed, there have been few other examples of autonomous reasoning tools with comparable impact and pervasiveness in practice.
  • Fair division has already proven a successful testbed for the application of theoretical work, thanks for the recently launched Spliddit webpage, which provides a user-friendly implementation for a number of algorithms in this field. This experience poses a number of questions and challenges for application-oriented research in fair division and beyond, such as data collection and analysis, possibly leading to new theoretical problems.
  • Interactive democracy comprises a variety of approaches to make democratic processes more engaging and responsive. For instance, successful design and implementation of online decision platforms presents a multidisciplinary research challenge.
  • Real electoral systems often have features that are absent in the single or multi-winner systems analyzed in textbooks and scientific papers. Voting theory and computational methods can help to identify non-monotonicity problems of real electoral systems, to provide normative benchmarks for institutional design, and to conduct influence and performance comparisons of different voting arrangements.

The Dagstuhl Seminar 19381 "Application-Oriented Computational Social Choice" brought together 46 invited participants of 15 different nationalities from 4 different continents, with three additional participants choosing to attend our seminar before participating to the Heidelberg Laureate Forum. The list of participants included researchers in Computer Science, Economics, and Political Science, three researchers from the industry (Microsoft, IBM, WinSet Group), and a lab technician.

For each of the focus topics described above, a 1-hour survey was prepared by one of the participants, obtaining an up-to-date overview of current research in the field and its main open problems. Each survey was scheduled on a different day, with 26 regular talks by participants complementing them in the program. Two rump sessions at the beginning of the week allowed a number of the participants to present recent findings, open problems and on-going research in a quick and informal way, stimulating the discussion for the rest of the week.

Given the focus of the seminar on application-oriented research, a special session was dedicated to the presentation of software developed by researchers participating to the seminar. Voting platforms were presented (Whale https://whale.imag.fr/ and OPRA https://opra.cs.rpi.edu/polls/main), a library for preference data (Preflib http://www.preflib.org/), a platform for online deliberation and consensus building (Vilfredo https://www.vilfredo.org/), as well as a number of tools to support experimental research in social choice. Moreover, the seminar hosted three live voting experiments during the week, two of which used a mobile experimental laboratory that was brought to Dagstuhl thanks to French CNRS and the help of a lab technician from University of Rennes. A detailed report of the experiments and an abstract of all the talks can be found below.

At the beginning of the week short sessions were reserved for individual self-introductions and for the proposition of potential group work. The organisers chose not to organize groups in advance, but to let them form in an iterative fashion during the seminar. A number of proposals were first made, then discussed and adapted, before participants signed up for specific group sessions. A total of 6 hours during the week was dedicated to group works, which led to significant advancements - a detailed report can be read below.

Overall, judging both from anecdotal personal feedback as well as the official results from the anonymous "Survey for Dagstuhl Seminar 19381" (with a median score of 10 out of 11 on the summary question "All in all, how do you rate the scientific quality of the seminar?" and similarly positive answers on the mix of participants, working atmosphere, etc.), the seminar was a very successful experience. It stimulated an already thriving research field to explore more applied research topics and scout for real-world problems. It allowed researchers to get first hand experience on how to run voting experiments, either on an Internet voting platform or in a laboratory, and allowed them to share their research practices. The work conducted in the groups was overall fruitful, already resulting in some paper drafts under preparation. The few suggestions for improvements mostly related to further broadening the mix of participants (more PhD students and junior researchers, more colleagues from nearby fields) and having a slightly less dense program (shorter talks, more time for work in small groups or unplanned activities).

The organisers wish to thank all the Dagstuhl staff for their professional support, the participants of the seminar for their positive attitude and enthusiasm, and the two collectors for putting together the abstracts that compose this report.

Summary text license
  Creative Commons BY 3.0 Unported license
  Umberto Grandi, Stefan Napel, Rolf Niedermeier, and Kristen Brent Venable

Classification

  • Artificial Intelligence / Robotics
  • Data Structures / Algorithms / Complexity

Keywords

  • Social Choice
  • Multi-Agent Systems
  • AI for the Social Good
  • Collective Decision Making

Documentation

In the series Dagstuhl Reports each Dagstuhl Seminar and Dagstuhl Perspectives Workshop is documented. The seminar organizers, in cooperation with the collector, prepare a report that includes contributions from the participants' talks together with a summary of the seminar.

 

Download overview leaflet (PDF).

Publications

Furthermore, a comprehensive peer-reviewed collection of research papers can be published in the series Dagstuhl Follow-Ups.

Dagstuhl's Impact

Please inform us when a publication was published as a result from your seminar. These publications are listed in the category Dagstuhl's Impact and are presented on a special shelf on the ground floor of the library.