Dagstuhl-Seminar 15131
Normative Multi-Agent Systems
( 22. Mar – 27. Mar, 2015 )
Permalink
Organisatoren
- Amit K. Chopra (Lancaster University, GB)
- Leon van der Torre (University of Luxembourg, LU)
- Harko Verhagen (Stockholm University, SE)
- Serena Villata (INRIA Sophia Antipolis - Méditerranée, FR)
Kontakt
- Susanne Bach-Bernhard (für administrative Fragen)
Programm
Normative systems are systems in the behavior of which norms play a role and which need normative concepts in order to be described or specified. A normative multi-agent system combines models for normative systems (dealing for example with obligations, permissions and prohibitions) with models for multi-agent systems. They provide a promising model for human and artificial agent coordination because they integrate norms and individual intelligence. They are a prime example of the use of sociological theories in multi-agent systems, and therefore of the relation between agent theory -- both multi-agent systems and autonomous agents -- and the social sciences -- sociology, philosophy, economics, legal science, etc. The seventh NorMAS seminar will feature challenging themes in broader computing and software engineering, namely social computing, governance, and agreement technologies. These themes are highly interdisciplinary, bringing together research strands from computing, information sciences, economics, sociology, and psychology. Further there is considerable excitement about these areas in academia, industry, and public policy organizations.
Social computing, as it is widely recognized today has its origins in Web 2.0 applications and had led to considerable government and industry push in Big Data. What is still lacking is a deeper conceptual understanding of the structure of social computing as it relates to multiple individuals and organizations. Also, the pragmatic aspects of information in social computing, such as social meaning, ownership, flow, and use, have received little systematic attention in social computing research so far. Norms explain both the social structure and the meaning of information, and therefore, have the potential to serve as a broad foundation for social computing.
Governance has come to the fore in discussions on public policy in the wake of disasters as the Deepwater Horizon oil spill and the Fukushima nuclear catastrophe. It refers to the administration of a socio-technical system (such as smart cities, smart grids, healthcare) by its stakeholders. Unlike technical kinds of administration, the administration of a socio-technical system has as its target the social expectations that stakeholders in the system would have of each other. This includes the elicitation of the social expectations, specifying systems in line with those expectations, monitoring their satisfaction and violation, and evolving the system to take into account potentially revised expectations.
Agreement technologies refer to computer systems in which autonomous software agents negotiate with one another, typically on behalf of humans, in order to come to mutually acceptable agreements. This theme aims to discuss the role of normative systems in agreement technologies, and the interaction of norms with other technologies such as ontologies, argumentation and trust management systems.
The three themes have overlapping topics, but their emphases show substantial differences. The social computing agenda has so far been driven by industry and applications, especially in social networks and social media. The governance agenda is driven largely by policy organizations (such as for health, safety, urban planning) with the aim of promoting better services for their citizens. Agreement technologies may be viewed as enabling both social computing and governance at one level and at another level having applications of its own such as consensus and auctions supported by specialized knowledge and infrastructure, such as required for argumentation, voting, negotiation, consensus-building, and vote accounting.
The multi-disciplinary workshop on Normative Multi Agents attracted leading international scholars from different research fields (e.g., theoretical computer science, programming languages, cognitive sciences and social sciences). The workshop was organized as follows: the organizers identified three relevant themes of research covering a wide and comprehensive spectrum of topics in the filed of Normative Agents, namely Social Computing, Governance, and Agreement Technologies. In the months preceding the workshop the chairs collected material from the participants. During the first day each participant present herself to the audience, and the chairs introduced the goal of the seminar, i.e., writing an handbook of Normative Multiagent Systems based on the roadmap produced during the previous edition of the Seminar, and the discussions during the current one. The participants were divided in groups corresponding to the areas identified as relevant in the field of Normative Multiagent Systems. Four invited talks have been proposed by scholars from different areas in the field, targeting in particular the three themes of the Seminar and an overview about Normative Multiagent Systems. The format was well received by the participants and conducive to discussion. It gave them the opportunity to give very focused presentations while keeping the audience attention. During the morning sessions, we started with an invited talk and we continued with short presentations by the Seminar participants about their personal contribution to Normative Multi-Agents (plus some time for QA). The afternoon sessions, other the contrary, were dedicated to group work and group discussions. The aim of these sessions was to build consensus material of the specific topics and to identify fundamental research directions. The material is expected to be refined and to be articulated in chapters intended as a first step for the development for the handbook for this emerging area of computer-science with close interactions with other disciplines.
Results
During the seminar, participants split in different working groups, centered around discussion themes relevant to NorMAS. Each working group was further divided into smaller working groups, each of which worked on specific topics. In the following paragraphs there is a summary of the discussion held by each working group.
Logic and reasoning. This theme included subgroups on topics such as deontic logic, argumentation, computation approaches, motivational attitudes, social games, and emotions.
Modeling. This theme included subgroups on issues such as taxonomies, law, conflicts, and norm dynamics.
Engineering. This theme included subgroups on themes such as interactions, agent programming, agent architecture, data-driven norms, institutions and technology, and reference architectures.
Simulation. This theme discussed issues of simulating multiagent systems to understand norm dynamics such as emergence and diffusion.
Applications. This theme included subgroups on killer applications for norms. Identified applications included governance, audit control, cybersecurity, jurisinformatics, and sociotechnical systems.
Each subgroup presented its findings twice to the entire seminar. Each subgroup identified past work, connections to other subgroups, and future work. Based on their presentations, we decided that each subgroup should write a chapter on its topic. This chapter will become part of a Handbook of Normative Multiagent Systems. This is in line with the roadmap produced during the previous edition of the Seminar and the discussions held during the present Seminar. The handbook will be an authoritative and detailed introduction for anyone seeking information on normative multiagent systems. The handbook will give a historical overview, present a survey of established techniques and open challenges, and discuss applications and directions. Our aim is to have to handbook sent for publication in a year's time. We already have a publisher lined up (College Publications).
- Huib Aldewereld (TU Delft, NL) [dblp]
- Diego Agustin Ambrossio (University of Luxembourg, LU) [dblp]
- Matteo Baldoni (University of Turin, IT) [dblp]
- Simon Caton (KIT - Karlsruher Institut für Technologie, DE) [dblp]
- Amit K. Chopra (Lancaster University, GB) [dblp]
- Rob Christiaanse (TU Delft, NL) [dblp]
- Silvano Colombo Tosatto (University of Luxembourg, LU) [dblp]
- Célia da Costa Pereira (University of Nice, FR) [dblp]
- Christoph Dorn (TU Wien, AT) [dblp]
- Hein Duijf (Utrecht University, NL) [dblp]
- Corinna Elsenbroich (University of Surrey - Guildford, GB) [dblp]
- Dov M. Gabbay (King's College London, GB) [dblp]
- Aditya K. Ghose (University of Wollongong, AU) [dblp]
- Guido Governatori (NICTA - Brisbane, AU) [dblp]
- Joris Hulstijn (TU Delft, NL) [dblp]
- Llio Humphreys (University of Turin, IT) [dblp]
- Franziska Klügl (University of Örebro, SE) [dblp]
- Ho-Pun Lam (NICTA - Brisbane, AU) [dblp]
- Beishui Liao (Zhejiang University, CN) [dblp]
- Daniel Moldt (Universität Hamburg, DE) [dblp]
- Robert Muthuri (University of Turin, IT) [dblp]
- Luis Gustavo Nardin (LABSS - ISTC - CNR - Rome, IT) [dblp]
- Martin Neumann (Universität Koblenz-Landau, DE) [dblp]
- Pablo Noriega (IIIA - CSIC - Barcelona, ES) [dblp]
- Julian Padget (University of Bath, GB) [dblp]
- Adrian Paschke (FU Berlin, DE) [dblp]
- Gabriella Pigozzi (University Paris-Dauphine, FR) [dblp]
- Axel Polleres (Wirtschaftsuniversität Wien, AT) [dblp]
- Livio Robaldo (University of Turin, IT) [dblp]
- Victor Rodriguez Doncel (Technical University of Madrid, ES) [dblp]
- Antonino Rotolo (University of Bologna, IT) [dblp]
- Ken Satoh (National Institute of Informatics - Tokyo, JP) [dblp]
- Judith Simon (IT University of Copenhagen, DK) [dblp]
- Munindar P. Singh (North Carolina State University - Raleigh, US) [dblp]
- Xin Sun (University of Luxembourg, LU) [dblp]
- Viviane Torres da Silva (IBM Research - Rio de Janeiro, BR) [dblp]
- Leon van der Torre (University of Luxembourg, LU) [dblp]
- Wamberto Vasconcelos (University of Aberdeen, GB) [dblp]
- Harko Verhagen (Stockholm University, SE) [dblp]
Verwandte Seminare
- Dagstuhl-Seminar 07122: Normative Multi-agent Systems (2007-03-18 - 2007-03-23) (Details)
- Dagstuhl-Seminar 09121: Normative Multi-Agent Systems (2009-03-15 - 2009-03-20) (Details)
- Dagstuhl-Seminar 12111: Normative Multi-Agent Systems (2012-03-11 - 2012-03-16) (Details)
- Dagstuhl-Seminar 18171: Normative Multi-Agent Systems (2018-04-22 - 2018-04-27) (Details)
Klassifikation
- artificial intelligence / robotics
- modelling / simulation
- society / human-computer interaction
Schlagworte
- norms
- agents
- social computing
- agreement technology
- governance