Marco Voss

A specification of the Agent Reputation and Trust (ART) testbed: experimentation and competition for trust in agent societies

A diverse collection of trust-modeling algorithms for multi-agent systems has been developed in recent years, resulting in significant breadth-wise growth without unified direction or benchmarks. Based on enthusiastic response from the agent trust community, the Agent Reputation and Trust (ART) Testbed initiative has been launched, charged with the task of establishing a testbed for agent trust- and reputation-related technologies. This testbed serves in two roles: (1) as a competition forum in which researchers can compare their technologies against objective metrics, and (2) as a suite of tools with flexible parameters, allowing researchers to perform customizable, easily-repeatable experiments. This paper first enumerates trust research objectives to be addressed in the testbed and desirable testbed characteristics, then presents a competition testbed specification that is justified according to these requirements. In the testbeds artwork appraisal domain, agents, who valuate paintings for clients, may gather opinions from other agents to produce accurate appraisals. The testbeds implementation architecture is discussed briefly, as well.

A Privacy Preserving Reputation System for Mobile Information Dissemination Networks

In a mobile information dissemination network mobile users, equipped with wireless devices, exchange information in a spontaneous manner whenever they come into communication range. Users have to specify what kind of information they are looking for and what kind of information they can offer. A priori there is no relation between users, literally spoken, they don’t know each other and confidence in newly collected information might be low. This work presents two reputation schemes, a simple and an extended version, for mobile information dissemination networks that, based on user ratings, increase a user’s confidence in some information source. As reputation systems collect sensitive personal information and monitor users’ behavior, privacy is an essential requirement — especially in a mobile scenario — that is neglected by many existing approaches. Using cryptographic group signatures and the concept of an observer, our extended reputation scheme guarantees high user privacy.

Privacy Preserving Online Reputation Systems

Reputation systems evolve as a mechanism to build trust in dynamic electronic societies. However, they are also a danger to privacy because they monitor a user’ s behavior. At the same time reputation systems offer the possibility to limit the information a user has to give away during a transaction to ensure accountablity.Privacy preserving reputation systems solve the conflict between anonymity and accountability. This paper examines privacy problems of current reputation systems and classifies them with respect to the location of stored information. Requirements for reputation systems that provide privacy protection are derived from this analysis. As result a new privacy preserving online reputation system is presented that uses locally stored coins to represent reputation information.

Monitoring component interaction in Jini federations

Jini is an infrastructure for spontaneous ad hoc service networks. It allows clients to find services without prior knowledge of their network surroundings. For service interaction proxy objects are used which are supplied by service providers. These proxy objects interact directly with the service provider. Compared to architectures that use a virtually central communications broker (like a CORBA ORB or an e-speak Core), this method offers a large amount of flexibility in the selection of an appropriate communication protocol. On the downside, debugging a distributed application using this approach is rather hard, as the interactions between clients and servers are not visible. This paper describes an approach using Javas dynamic proxies that allows component interaction in a Jini federation to be traced. By putting the functionality into the Jini lookup service, the approach is generic and transparent for both services and clients.

Simulating data dissemination techniques for local reputation systems

In distributed scenarios the robustness of a reputation mechanism depends on the data available for computation. Especially in ad-hoc networks the amount of available data is restricted due to communication latency. In this paper we present a formal model for describing reputation systems with limited information and evaluate different data dissemination techniques for supporting the computation of reputation in dynamic multi-agent systems by simulation.