(WORKSHOP CO-FUNDED BY THE EUROPEAN SCIENCE FOUNDATION, ESF)
Agreement is one of the crucial social concepts that helps human agents to cope with their social environment and is present in all human interactions. In fact, without agreement there is no cooperation and ultimately social systems cannot emerge. Agreement is necessary in our everyday life.
Until recently, the concept of agreement was a domain of study mainly for philosophers, sociologists and was only applicable only to human societies. In the last thirty years, the growth of disciplines such as social psychology, sociobiology, social neuroscience, and the development of adaptive multi-agent artificial systems (MAS), together with the spectacular emergence of the information society technologies, have changed this situation. Presently, agreement and all the processes and mechanisms implicated in reaching agreements between different kind of agents are a subject of research and analysis with very many perspectives.
Provided that all these approaches are relevant for a robust understanding and an efficient implementation of artificial social systems, the dialogue and knowledge transfer among these new disciplines appears as a must in social-oriented science and technology.
Indeed, the scientific research in the area of agreement mechanisms for virtual societies is a recent discipline oriented to increase the reliability and performance of electronic communities by introducing in such communities these well known human social control mechanisms. Computer science has moved from the paradigm of an isolated machine to the paradigm of a network of systems and of distributed computing. Likewise, artificial intelligence is quickly moving from the paradigm of an isolated and non-situated intelligence to the paradigm of situated, social and collective intelligence.
Agent Technology is the latest paradigm of software engineering methodology. The development of autonomous, mobile, and intelligent agents brings new challenges to the field. Agent technologies and multi-agent-systems are one of the most vibrant and active research areas of computer science. At the same time commercial applications of agents are gaining attention. The construction of artificial (agent) societies leads to questions that already have been asked for human societies. Computer Scientists have adopted terms like emerging behavior, self-organization, and evolutionary theory in an intuitive manner. Multi-agent-system researchers have started to develop agents with "social" abilities and complex "social" systems. However, most of these systems lack the foundation of the social sciences.
The wide range of social theories offers many different solutions to problems found in complex (computer) systems. Which theories, to apply how and when is a major challenge. In developing agents and multi-agent-systems computer scientists have used sociological terms like negotiation, reputation, trust, interaction, contracts, organisation, cohesion, social order, or collaboration.
Meanwhile an interdisciplinary area called Socionics, the bridge between sociology and computer science, is beginning to establish itself. The realisation that the behaviour of societies cannot fully be explained by macro-theories only, and the progress made in agent technology opened the way to new models of societies in which both macro-theories and micro-theories are incorporated. However, the integration of these theories is still insufficient. The development of the socionics research area and the increased interest in dynamics of behaviour and structures of or within agents in hybrid organisations requires the investigation of new modelling concepts.
When looking at the actually implemented systems many difficult challenges have to be solved with respect to the behaviour and the relationships within the involved and implemented entities of a system. Especially the introduction of norms and institutions in the social world has helped to organize our world. The transfer of these concepts into the field of MAS seems to be a promissing direction. At the same time the complete understanding of establishment of norms and institutions still needs research. Most interesting is that MAS are now seen as a good vehicle in terms of metaphers, concepts, and tools to support this research direction.
Social networks have been a good tool to provide an effective tool for sociologists studying individual behaviors in a complex social system and testbed for the study and evaluation of artificial agent societies. The mutual dependencies between organizations and their personal / artifical actors is a demanding testbed for theories and applications of MAS. The future will provide more and more connected and mutually dependent organizations / people / artificial agents. The consequences still have not been understood.
The way to approach all these challenges can be from theoretical, experimental, empirical, prototypical, and applied work.
Trust is a critical prerequisite of any agreement process. Without trust, agents could set up any constructive dialogue nor work cooperatively together. Trust helps to reduce the complexity of decisions that have to be taken in the presence of many risks. Similarly, reputation is a universal concept that has been present in human societies for a long time. Reputation is one of the most relevant elements that we use to build trust in others.
Most agent models assume secure and reliable communication to exist between agents. However, this ideal situation is seldom met in reality. In fact, many techniques (e.g. contracts, signatures, long-term personnel relationships, reputation) have been evolved over time to detect and prevent deception and fraud in human communication, exchanges and relations, and hence to assure trust between agents. Artificial societies will need analogous techniques.
Trust is more than secure communication, e.g., via public key cryptography techniques. For example, the reliability of information about the status of your trade partner has little to do with secure communication. With the growing impact of electronic societies, trust and privacy become more and more important. Trust is important in applications such as human-computer interaction to model the relationship between users and their personal assistants. Different kinds of trust are needed: trust in the environment and in the infrastructure (the socio-technical system) including trust in your personal agent and in other mediating agents; trust in the potential partners; trust in the warrantors and authorities (if any). Another growing trend is the use of reputation mechanisms, and in particular the interesting link between trust and reputation. Many computational and theoretical models and approaches to reputation have been developed in the last years.
Trust appears to be foundational for the notion of "agency" and for its defining relation of acting "on behalf of". It is also critical for modeling and supporting groups and teams, organizations, co-ordination, negotiation, with the related trade-off between individual utility and collective interest; or in modeling distributed knowledge and its circulation. In several cases the electronic medium seems to weaken the usual bonds in social control: and the habit or disposition to cheat grow stronger. In experiments of cooperation supported by computers it has been found that people are more leaning to defeat than in face-to-face interaction, and a preliminary direct acquaintance reduces this effect. So, computer technology can even break trust relationships already held in human organizations and relations, and favor additional problems of deception and trust.
Yet, not only there is not a shared and dominant model of trust and reputation as mental attitude, as decision and action, and as a social relationship. What is needed is a general and principled theory of trust, of its cognitive and affective components, and of its social functions.
CONVENORS:
Ramon L. de Mántaras,
Professor at the Artificial Intelligence Research Institute (IIIA) of the Spanish Research Council (CSIC). SPAIN
Ramón López de Mántaras is a pioneer of Artificial Intelligence in Spain, with contributions, since 1976, in the following areas: Unsupervised Learning & Clustering, Pattern Classification, Tactile Recognition in Robotics, Multiple-valued and Fuzzy Logics for Approximate Reasoning, Knowledge Acquisition, Medical Expert Systems, Inductive Learning, Case-Based Reasoning, Autonomous Robots, AI & Music, and Bayesian Learning.
Address:
IIIA-CSIC
Campus Universitat Autonoma de Barcelona
08193 Bellaterra, Catalonia,
Spain
+34-93-580 95 70
mantaras(at)iiia(dot)csic(dot)es
Carles Sierra, Professor at the Artificial Intelligence Research Institute (IIIA) of the Spanish Research Council (CSIC). SPAIN
Carles Sierra’s research focuses on electronic institution related tools, semantic web and electronic institutions, and electronic Reputation methods.
Address:
IIIA-CSIC
Campus UAB,
08193 Cerdanyola, Catalonia, Spain
+34-93-580 95 70
sierra at iiia.csic.es
Oscar Vilarroya,
Director of the Chair “The Social Brain”, Autonomous University of Barcelona, SPAIN
The cognitive scientist Oscar Vilarroya is involved in neuroimaging related to mental disorders, normal behaviour, and biological mechanism underlying social behaviour. He has developed theories emphasising the embodiment of intelligence.
Address:
Universitat Autònoma de Barcelona
Centre Fòrum. Hospital del Mar.
Carrer Llull 410
08019 Barcelona
SPAIN
+ 34 93 254 13 13
Oscar.vilarroya@uab.cat
