Conference Report: The Twelfth International Workshop on Computational Logics in Multi-Agent Systems

Multi-agent systems are systems of interacting autonomous agents or components that can perceive and act upon their environment to achieve their individual goals as well as joint goals. Research on such systems integrates many technologies and concepts in artificial intelligence and other areas of computing as well as other disciplines. Over recent years, the agent paradigm gained popularity, due to its applicability to a full spectrum of domains, from search engines to educational aids to electronic commerce and trade, e-procurement, recommendation systems, simulation and routing, to mention only some.

Computational logic provides a well-defined, general, and rigorous framework for studying syntax, semantics and procedures for various tasks by individual agents, as well as interaction amongst agents in multi-agent systems, for implementations, environments, tools, and standards, and for linking together specification and verification of properties of individual agents and multi-agent systems.

The series of workshops on Computational Logics in Multi-Agent Systems –  CLIMA – aims at providing a forum for discussing techniques, based on computational logic, for representing, programming and reasoning about agents and multi-agent systems in a formal way.

The twelfth edition of CLIMA took place between the 17th and 18th of July 2011 in Barcelona and was a great success. Besides regular sessions the workshop featured special sessions on Logics for Games and Social Choice, organized by Thomas Ågotnes, and Norms and Normative Multi-Agent Systems, organised by Guido Boella and Leon van der Torre.

The workshop opened with a session on Secrets and Trust. In his invited talk, Simon Parsons discussed the use of argumentation for reasoning about which individuals to trust, and for relating sources of information to conclusions drawn from information provided by those sources. Robert Demolombe then looked at information and trust propagation, exploring formal notions of sincerity, competence, vigilance, cooperation, validity and completeness. Afterwards, Sara Miner More and Pavel Naumov presented a follow-up of their CLIMA XI work, with a theoretical study of a logic of dependence between secrets, addressing the question: which secrets functionally determine which others?

In the second half of the morning, we had three presentations on Knowledge and Beliefs, covering topics such as knowledge-based protocols, security in information exchange, but also modalities for modeling beliefs and information sources, belief merging, information aggregation, and the concept of definability. Hans Van Ditmarsch and Fernando Soler-Toscano opened the session by presenting a three-step protocol which allows two players to (publicly) inform each other about their playing cards without making it known to a third player, the eavesdropper. Emiliano Lorini, Laurent Perrussel, and Jean-Marc Thévenin then illustrated a framework for processing signed information before incorporating it in the agent’s beliefs. The last presentation of the morning, based on work by Hans Van Ditmarsch, David Fernández and Wiebe van der Hoek, dealt with epistemic logic: under what conditions can individual Kripke models be uniquely characterised (up to bi-simulation) by a single epistemic formula?

The afternoon, entirely devoted to the special session on Logics for Games and Social Choice, was opened by Ulle Endriss‘s invited talk. While illustrating recent work conducted by members of his group at the University of Amsterdam, Ulle showed many different ways in which modern logic can contribute to the study of social choice theory. Following the invited talk, five regular papers covered theoretical aspects and applications. Jan Van Eijck presented a new (geometric) proof of the Gibbard-Satterthwaite theorem, based on the Saari triangle for three alternatives, and  elaborated on the notion of non-manipulability and the need for a finer granularity. Tiago De Lima‘s talk on alternating-time announcement logic presented a very general logic of action and change. Jan Calta, Dmitry Shkatov and Holger Schlingloff presented results on synthesising strategies for multi agent systems. Thomas Ågotnes and Natasha Alechina then discussed how coalition logic reasoning can be done in standard PDL-like logics. A final presentation by Daniele Porello and Ulle Endriss illustrated a proposal to regard ontology merging as a problem of social choice.

The second day was opened by a session featuring four talks on Cooperation, covering aspects related to interaction protocols, teams, commitments, query-answering, monitoring, verification and diagnosis. Özgür Kafali and Paolo Torroni started with a systematic analysis of types of delegation, similarity delegation and improper delegation. Taolue Chen, Marta Kwiatkowska, David Parker and Aistis Simaitis proposed to use probabilistic model checking as an analysis tool for organisational construction. Samy Sá and João Alcântara presented some strategies to generate cooperative answers in query answering systems: when there is no correct answer to a given query, it is more helpful to return some answers related to the query using query relaxation. In the last presentation, Özgür Kafali, Francesca Toni and Paolo Torroni discussed an assumption-based argumentation approach to diagnosis of commitment exceptions.

Three presentations on Logic and Languages for agent programming concluded the morning. Domenico Corapi, Daniel Sykes, Katsumi Inoue and Alessandra Russo discussed a proposal for rule learning, aimed to enrich abductive reasoning with a probabilistic component as well to model inductive logic programming tasks as a special form of abductive reasoning. Richard Stocker, Maarten Sierhuis, Louise Dennis, Clare Dixon and Michael Fisher then presented a formal semantics for the Brahms modelling and simulation framework for human-agent teamwork. Finally, Alfredo Gabaldon showed how norm-enforcing mechanisms can be accommodated in the Golog situation calculus-based programming language.

The final part of the CLIMA programme featured the special session on  Norms and Normative Multi-Agent Systems, with one invited talk and five regular paper presentations. In his invited talk, Jan Broersen discussed modeling of obligations to attempt an action in a probabilistic stit framework extended with deontic modalities, and the effects of reasoning with probabilities on the semantics of deontic modalities like obligation and prohibition. Jan’s talk was followed by two presentations on actions and norms that used modal style logics. Andreas Herzig, Emiliano Lorini and Nicolas Troquard defined a logic of action which enables reasoning about the distinction between physical actions bringing about brute facts and institutional actions bringing about institutional facts. Mathieu Beirlaen and Christian Strasser, instead, proposed a paraconsistent approach for dealing with normative conflicts in multi-agent systems. Afterwards, a last group of three presentations discussed research in norms and normative multi-agent systems using rules style logics. Marco Alberti, Ana Sofia Gomes, Ricardo Gonçalves, João Leite and Martin Slota showed an application of a logic combining DL ontologies with rule-based knowledge bases for representing and reasoning about norms in multi-agent systems. Nir Oren, Wamberto Vasconcelos, Felipe Meneguzzi and Michael Luck then presented an approach in which norms are constraints and are used to generate plans that satisfy those norms that yield the highest utility. In the concluding talk, Guido Governatori and Antonino Rotolo extended a logic of violation with time, thus enabling the representation of compliance with respect to different types of legal obligation and different temporal constraints over them, as well as the representation of reparative obligations.

The 22 papers presented at CLIMA XII and published in [1] were selected from 43 submissions. These were on average of very high quality. In line with the high standards of previous editions, the final acceptance rate, after two rounds of reviewing and selection, was circa 50%. The programme committee consisted of 48 top-level researchers from 35 institutions located in 5 continents and 15 countries. 9 additional reviewers helped in the process.

General information about the workshop series, with links to past and future events, can be found on the CLIMA workshop series home page, http://centria.di.fct.unl.pt/~clima/.

[1] João Leite, Paolo Torroni, Thomas Ågotnes, Guido Boella, and Leon van der Torre, editors.
Computational Logic in Multi-Agent Systems – 12th International Workshop, CLIMA XII, Barcelona, Spain, July 17-18, 2011. Proceedings, volume 6814 of Lecture Notes in Computer Science. Springer, 2011.