Towards the Next Generation of Computational Trust and Reputation Models

Abstract

The scientific research in the area of computational trust and reputation 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 a situated, social and collective intelligence. This new "social" dimension is the main responsible of the increasing interest on computational trust and reputation mechanisms applied to electronic societies.Computational trust and reputation systems have been recognized as key factors for successful electronic commerce adoption. These systems are used by intelligent software agents both as a mechanism of search for trustworthy exchange partners and as an incentive in decision-making about whether or not to honour contracts. Reputation is also used in electronic markets as a trust-enforcing, deterrent, and incentive mechanism to avoid cheaters and frauds. Another area of application in agent technology is teamwork and cooperation.However, in spite of the obvious utility of using a trust and reputation system, they are still not a usual element you can find in an agent architecture like it is the case for example of planners or communication modules. Trust and reputation systems still belong to the set of "not so important" elements of an agent architecture. This is something it has to change if we really want artificial socially intelligent entities.If we analyze which computational trust and reputation models are used nowadays, we will notice that only the simplest models are really used (see for example e-Bay[1]). Till now, the use of complex trust and reputation models has been somehow questioned and reduced to academic environments. Why an artificial agent has to use one of these complex models if it is evolving in a simple community with 10 or 15 agents that have limited interaction capabilities? In other words, the applications and environments are too simple to justify a complex trust and reputation model that only has sense in a socially complex environment.However this situation is changing very quickly. The increase of the global connectivity (everything is inter-connected anytime and everywhere) is bringing us a new concept of environment where virtual entities need to interact among

Notorious Places: Image, Reputation, Stigma. The Role of Newspapers in Area Reputations for Social Housing Estates

“…The application of concepts of reputation to market transactions has taken another step forward since the 1990s as a result of online marketplaces, so that one of the largest literatures we found on reputation was (surprisingly to us) that on distributed artificial intelligence (DAI) systems in computing science (Sabater-Mir, 2006). In online market places [a type of multi-agent system (MAS)], self-interested agents with limited information, limited capabilities, and most-importantly, without face-to-face contact, interact for mutual benefit.…”

Asset Management System through the design of a Jadex Agent System

“…This ability is just estimated in such a way similar to the reputation of agents in models of trust applied to Agent systems (Josang et al, 2006;Sabater-Mir and Sierra, 2005).…”

Privacy Protection in Trust Models for Agent Societies

“…Therefore we assume (as many researchers in Distributed AI) that each member of a society of agents is in charge of computing the reputation of all other agents that belong to this society. Many trust models been proposed and they are very different among them [10]. But most of them consider direct experiences and witness information as the main information sources.…”