12. – 15. Mai 2014, Dagstuhl-Seminar 14202

JA4AI – Judgment Aggregation for Artificial Intelligence


Franz Dietrich (CNRS, FR & University of East Anglia, GB)
Ulle Endriss (University of Amsterdam, NL)
Davide Grossi (University of Liverpool, GB)
Gabriella Pigozzi (University Paris-Dauphine, FR)


Marija Slavkovik (University of Bergen, NO)

Auskunft zu diesem Dagstuhl-Seminar erteilt

Dagstuhl Service Team


Dagstuhl Report, Volume 4, Issue 5 Dagstuhl Report
Dagstuhl's Impact: Dokumente verfügbar
Programm des Dagstuhl-Seminars [pdf]


Judgment aggregation is a group decision-making theory, developed in the last decade, that studies how to reach group decisions on logically interconnected issues by aggregation of individual decisions on those issues. The interest of computer science in group reasoning and decision-making theories is driven by the increase of distribution of information and computation as features of various Internet-based services that dominate the information technology market.

Judgment aggregation studies collective decision-making as a process whereby individual opinions concerning the acceptance or rejection of a set of issues are aggregated into one collective judgment. The problem is for the aggregation process to preserve, in a non-trivial way, some 'rational' aspects of the individual to-be-aggregated stances like, in particular, logical consistency. A wealth of results have highlighted how the rationality of a collective decision may clash with other desirable properties of a process of aggregation one may wish to require (e.g., anonymity of the voters, independence of the aggregated issues, to mention a few).

Judgment aggregation research, originally studied in law, was propelled into other disciplines with its establishment as a separate discipline from preference aggregation in the early 2000's. The first half of the decade was marked by studies of aggregation properties that cannot be jointly satisfied by one aggregation function, usually referred to as `impossibility results'. These studies were mostly conducted by researchers from political science, law, economics, mathematics, and philosophy. The second half of the decade witnessed an increase of interest in judgment aggregation of researchers from artificial intelligence (AI), specifically knowledge representation and reasoning (KR), and multi-agent systems (MAS).

Research on judgment aggregation, from the computer scientific perspective, has splintered in many directions, with scholars pursuing very different lines of research: judgment aggregation and logic, judgment aggregation and complexity theory, judgment aggregation and relations to preference aggregation, judgment aggregation and belief merging, judgment aggregation and argumentation, to mention a few. At the same time work in judgment aggregation has diversified in non-computer science disciplines: judgment aggregation and deliberation, judgment aggregation and strategic voting, judgment aggregation and probabilistic opinion pooling, to mention a few. Despite the common research thread, having so many disciplines involved make it difficult to keep track of the research advancements across all domains.

The goal of this Dagstuhl seminar was to give researchers across the contributing disciplines an integrated overview of the current research and interests in judgment aggregation and of its emerging trends, and by doing this, to kick-start a lasting interdisciplinary network bridging the computer science/humanities divide in the field. To accomplish this goal, we structured the seminar around four types of events:

  • Invited tutorials - three invited overview talks aimed to introduce the interdisciplinary audience to the origins and advancements of judgment aggregation in law, political science and computer science.
  • Contributed talks - fourteen contributed talks of thirty minutes each.
  • Networking sessions -- two free networking sessions.
  • Rump session -- open to all participants to present new ideas.

The topics of the invited talks were chosen so as to give a foundation of the disciplines in which judgment aggregation originated and was formalised, as well as to motivate the interest of judgment aggregation for computer science. Although we expected that all of the participants would be familiar with at least one of these foundational topics, we also expected them to be unfamiliar with at least one as well. The tutorial lectures aimed to homogenise the background knowledge in judgment aggregation among the participants.

The contributed talks aimed to introduce the community with the recent work of the speakers. We accommodated fourteen talks, possibly compromising on the length of the talk itself in the interest of allowing space for questions. We are happy to observe that there was a lively debate after each of the talks, which we expect shall contribute towards advancement of each of the presented works.

Given the short period of three days and prior Dagstuhl experience of the organisers, we decided to not structure the networking session and simply allow for a time for the participants to talk to each other and get to know about each other's work and interests. The enthusiastic discussions following the contributed talks typically continued into the networking sessions.

The rump session was free for a last-minute sign up to all participants. Each interested person was given a five-minute time slot to present an idea that emerged during the seminar or a work in progress. A third of the participants took this opportunity to present. This was a very lively and well received part of the seminar. In retrospect, a similar session would have been well received also at the beginning of the seminar, giving the participants more time to discuss the presented ideas.

Summary text license
  Creative Commons BY 3.0 Unported license
  Franz Dietrich and Ulle Endriss and Davide Grossi and Gabriella Pigozzi and Marija Slavkovik


  • Artificial Intelligence / Robotics
  • Semantics / Formal Methods
  • Verification / Logic


  • Judgment aggregation
  • Collective decision-making
  • Belief merging
  • Multi-agent systems
  • Logic


In der Reihe Dagstuhl Reports werden alle Dagstuhl-Seminare und Dagstuhl-Perspektiven-Workshops dokumentiert. Die Organisatoren stellen zusammen mit dem Collector des Seminars einen Bericht zusammen, der die Beiträge der Autoren zusammenfasst und um eine Zusammenfassung ergänzt.


Download Übersichtsflyer (PDF).

Dagstuhl's Impact

Bitte informieren Sie uns, wenn eine Veröffentlichung ausgehend von Ihrem Seminar entsteht. Derartige Veröffentlichungen werden von uns in der Rubrik Dagstuhl's Impact separat aufgelistet  und im Erdgeschoss der Bibliothek präsentiert.


Es besteht weiterhin die Möglichkeit, eine umfassende Kollektion begutachteter Arbeiten in der Reihe Dagstuhl Follow-Ups zu publizieren.