Dagstuhl Seminar 25401
Societal Impact of Computational Social Choice
( Sep 28 – Oct 02, 2025 )
(Click in the middle of the image to enlarge)
Permalink
Please use the following short url to reference this page:
https://www.dagstuhl.de/25401
Organizers
- Martin Lackner (TU Wien, AT)
- Nicholas Mattei (Tulane University - New Orleans, US)
- Arianna Novaro (Université Paris 1 Panthéon-Sorbonne, FR)
- Clemens Puppe (KIT - Karlsruher Institut für Technologie, DE)
Contact
- Marsha Kleinbauer (for scientific matters)
- Simone Schilke (for administrative matters)
Shared Documents
- Dagstuhl Materials Page (Use personal credentials as created in DOOR to log in)
Impacts
- Social Choice with Text : Collective Decision Making in the LLM Era - Grandi, Umberto; Bredereck, Robert; Delemazure, Théo; Maly, Jan; Mattei, Nicholas; Szufa, Stanislaw; Nardi, Oliviero; Maudet, Nicolas; Endriss, Ulle - HAL Inria, 2026. - 18 pp..
- Computational Social Choice : Research and Development : 25th International Conference on Autonomous Agents and Multiagent Systems AAMAS 2026 (Blue Sky Ideas Track) : accepted paper - Baumeister, Dorothea; Berker, Ratip Emin; Boehmer, Niclas; Bouveret, Sylvain; Darmann, Andreas; Novaro, Arianna; Mattei, Nicholas; Lang, Jerome; Lackner, Martin; Faliszewski, Piotr - Cornell University : arXiv.org, 2026. - 12 pp..
Schedule
The Dagstuhl Seminar 25401 on "Societal Impact of Computational Social Choice" gathered 41 participants, from 11 countries and 4 continents. Most of the participants' primary area of research was Computational Social Choice(COMSOC), an interdisciplinary field which brings together computer scientists, economists, mathematicians, philosophers and political scientists. The field is primarily concerned with designing algorithms and systems for collective decision-making, and analysing these algorithms and systems from an axiomatic, strategic, and computational perspective. Moreover, one of the seminar participants, Mathijs Kemp, brought a perspective from outside of academia, as he is the founder of the Parta platform for collective decision-making.
The seminar program focused on the three main topics of data, participation, and time in computational social choice, under the two overarching themes of domain restrictions and societal impact:
- Data has become an essential element for COMSOC research. In fact, by using open repositories and datasets (such as the PrefLib and PabuLib libraries), synthetic datasets, as well as tools (such as PrefPy, ABC Voting Rules and many others), COMSOC researchers can now implement and test their algorithms for collective decision-making. On the one hand, the results of these tests can inform recommendations on which methods actually perform better on desirable objectives in practice. On the other hand, they may highlight the presence of some underlying structure, which could lead to a more refined theoretical analysis.
- Participation has been formalized and studied in classical social choice theory by means of various mathematical axioms capturing the intuitive idea that a voting mechanism should incentivize voters to actually express their opinions. More recently, many municipalities and public institutions have moved towards various forms of participatory and digital democracy, with the goal of increasing the citizens' active role and engagement in the public life of their communities. Perhaps the main such example is that of participatory budgeting, where the citizens can propose and then vote on how to spend a percentage of a public budget - a practice that has been implemented in many cities around the world, and that has attracted the interest of COMSOC researchers. Moreover, in order to effectively elicit citizen participation, we also have to consider the (digital) platforms over which the collective decisions take place.
- Time is a dimension that has received increasing attention when studying problems of collective decision-making. Some examples of time-aware formalisms in COMSOC include: iterative voting, i.e., the interplay of strategic voting with time, as agents can modify their votes after observing the current results; perpetual voting, i.e., the analysis of fairness in elections over time; dynamic social choice, i.e., the change in agents' preferences, available resources, or number of participants over time. Indeed, numerous real-world problems include time as a component or have a repeated nature, e.g., the allocation of weekly chores among roommates.
Each of these three topics, as well as the two overarching themes, had a dedicated talk session in the seminar program (with the exception of the theme of societal impact which had two sessions), for a total of 18 talks. The talk sessions were structured as three short research talks followed by a panel discussion between the session speakers and the audience.
The seminar program also included 2 invited talks to report on the process and the interaction between policy-makers and academic advisors. The first one was by Sam Hirsh, who was involved in a legal process to reconfigure the electoral voting maps in the United States, and the second one was by Friedrich Pukelsheim, who was a member of the expert committee advising the German Bundestag in the process of the electoral reform of the German Bundeswahlgesetz.
Finally, the program included an experiment session by Théo Delemazure and Jérôme Lang, as well as numerous working group sessions and some time for informal discussions.
We wish to thank all of the participants, the invited speakers, our report collector Emin Berker, as well as the great Dagstuhl staff, for their valuable contribution to the success of the seminar.
Creative Commons BY 4.0
Martin Lackner, Nicholas Mattei, Arianna Novaro, and Clemens Puppe
Computational Social Choice (COMSOC) is an interdisciplinary field between classical social choice theory in economics and theoretical computer science. The focus is to study algorithms for collective decision-making problems, such as political elections, the allocation of resources, and so on.
In this Dagstuhl Seminar, we want to focus on three main topics. The first one is data, which has become an essential element for COMSOC research. In fact, thanks to the availability of open libraries, datasets and tools, researchers can now implement and test their algorithms for collective decision-making on real-life data, complementing their theoretical results. The second one is participation, as in recent years many municipalities and public institutions have moved towards various forms of participatory and digital democracy, with the goal of increasing the citizens' active role and engagement in the public life of their communities. The third one is time, as although many collective decision-making problems have an underlying repeated nature, this dimension has thus far not received the deserved attention within standard COMSOC models.
Moreover, we aim to address these topics under the two overarching themes of domain restrictions and societal impact: while domain restrictions can be seen as a methodological question over the input of our problems, societal impact can be seen as part of their output, i.e., the applications originating from theoretical research.
Martin Lackner, Nicholas Mattei, Arianna Novaro, and Clemens Puppe
Please log in to DOOR to see more details.
- Dorothea Baumeister (HS Bund f. Öffentl. Verwaltung - Brühl, DE) [dblp]
- Gerdus Benade (Boston University, US) [dblp]
- Ratip Emin Berker (Carnegie Mellon University - Pittsburgh, US)
- Niclas Boehmer (Hasso-Plattner-Institut, Universität Potsdam, DE) [dblp]
- Sylvain Bouveret (University of Grenoble, FR) [dblp]
- Florian Brandl (Universität Bonn, DE) [dblp]
- Felix Brandt (TU München - Garching, DE) [dblp]
- Robert Bredereck (TU Clausthal, DE) [dblp]
- Markus Brill (University of Warwick - Coventry, GB) [dblp]
- Martin Bullinger (University of Oxford, GB) [dblp]
- Andreas Darmann (Universität Graz, AT) [dblp]
- Théo Delemazure (University of Amsterdam, NL)
- Edith Elkind (Northwestern University - Evanston, US) [dblp]
- Ulle Endriss (University of Amsterdam, NL) [dblp]
- Piotr Faliszewski (AGH University of Science & Technology - Krakow, PL) [dblp]
- Paul Gölz (Cornell University - Ithaca, US) [dblp]
- Umberto Grandi (University Toulouse Capitole, FR) [dblp]
- Davide Grossi (University of Groningen, NL) [dblp]
- Mathijs Kemp (Vennster - Almere, NL)
- Sonja Kraiczy (University of Oxford, GB)
- Martin Lackner (TU Wien, AT) [dblp]
- Jérôme Lang (CNRS - Paris, FR) [dblp]
- Patrick Lederer (UNSW - Sydney, AU) [dblp]
- Jan Maly (Wirtschaftsuniversität Wien, AT)
- Nicholas Mattei (Tulane University - New Orleans, US) [dblp]
- Nicolas Maudet (Sorbonne University - Paris, FR) [dblp]
- Reshef Meir (Technion - Haifa, IL) [dblp]
- Oliviero Nardi (TU Wien, AT) [dblp]
- Arianna Novaro (Université Paris 1 Panthéon-Sorbonne, FR) [dblp]
- Dominik Peters (University Paris-Dauphine, FR) [dblp]
- Jannik Peters (National University of Singapore, SG) [dblp]
- Marcus Pivato (Université Paris 1 Panthéon-Sorbonne) [dblp]
- Friedrich Pukelsheim (Universität Augsburg, DE) [dblp]
- Clemens Puppe (KIT - Karlsruher Institut für Technologie, DE) [dblp]
- Ulrike Schmidt-Kraepelin (TU Eindhoven, NL) [dblp]
- Piotr Skowron (University of Warsaw, PL) [dblp]
- Arkadii Slinko (University of Auckland, NZ) [dblp]
- Stanislaw Szufa (CNRS - Paris, FR) [dblp]
- Frederik Van De Putte (Erasmus University - Rotterdam, NL) [dblp]
- Toby Walsh (UNSW - Sydney, AU) [dblp]
- Tomasz Was (University of Oxford, GB) [dblp]
Related Seminars
- Dagstuhl Seminar 28181: Models and Computation of Proportional Representation (2028-05-01 - 2028-05-05) (Details)
Classification
- Computer Science and Game Theory
- Multiagent Systems
Keywords
- computational social choice
- data
- time
- participation
