Trust is a crucial basis for interactions among parties in large, open systems. Yet, the scale and dynamism of such systems make it infeasible for each party to have a direct basis for trusting another party. For this reason, the participants in an open system must share information about trust. However, they should not automatically trust such shared information. This paper studies the problem of propagating trust in multiagent systems. It describes a new algebraic approach, shows some theoretical properties of it, and empirically evaluates it on two social network datasets. This evaluation incorporates a new methodology that involves dealing with opinions in an evidential setting. 1

Abstract. A trust metric is a technique for predicting how much a user of a social network might trust another user. This is especially beneficial in situations where most users are unknown to each other such as online communities. We think the recent tumultuous evolution of social networking demands for a collective research effort. With this in mind we created Trustlet.org, a platform consisting of a wiki for open research on trust metrics. The goal of Trustlet is to collect and distribute trust network datasets and trust metrics code as free software, in order to facilitate the comparison of different trust metrics algorithms and a more coherent progress in this field. At present we made available some social network datasets and code for some trust metrics. In this paper we also report a first empirical evaluation of different trust metrics on the Advogato social network dataset.

In activity-centric environments where people from different companies and disciplines work remotely together and where new virtual teams are formed and dissolved continuously, how to find the most suitable collaboration partner for a given task and how well one partner is able to collaborate with another one are challenging research questions. Determining and considering people’s professional competencies, collaboration behavior and relationships is a prerequisite to enhance the overall collaboration performance and success, because these factors highly impact on the notion of trust used to select and grade partners. In this paper we analyze these factors and their impact on trust relationships in modern service-oriented collaboration environments. We present VieTE, a framework for trust emergence therein supporting the analysis of trust between partners in various contexts and from different views. In contrast to other approaches, which mostly rely on manual and subjective user feedback, VieTE monitors automatically collaboration efforts and deduces trust between any two partners based on past collaboration, previous successes, and individual competencies. 1

An evidence-based account of trust is essential for an appropriate treatment of application-level interactions among autonomous and adaptive parties. Key examples include social networks and service-oriented computing. Existing approaches either ignore evidence or only partially address the challenges of mapping evidence to trustworthiness and combining trust reports from imperfectly trusted sources. This paper develops a mathematically well-formulated approach that naturally supports discounting and combining evidence-based trust reports. This paper understands an agent Alice’s trust in an agent Bob in terms of Alice’s certainty in her belief that Bob is trustworthy. Unlike previous approaches, this paper formulates certainty in terms of evidence based on a statistical measure defined over a probability distribution of the probability of positive outcomes. This definition supports important mathematical properties ensuring correct results despite conflicting evidence: (1) for a fixed amount of evidence, certainty increases as conflict in the evidence decreases and (2) for a fixed level of conflict, certainty increases as the amount of evidence increases. Moreover, despite a subtle definition of certainty, this paper (3) establishes a bijection between evidence and trust spaces, enabling robust combination of trust reports and (4) provides an efficient algorithm for computing this bijection.

Abstract. Trust networks are directed weighted graphs whose nodes represent agents and edges represent trust between agents. This paper proposes a trust-based recommendation approach, which can recommend trustworthy agents to a requester in a trust network. We consider a good recommendation as one to an agent that the requester’s trusted neighbors trust highly. We relate the recommendation problem to the graph similarity problem, and define the similarity measurement as a mutually reinforcing relation. By calculating the vertex similarity between the trust network and a structure graph (a path graph of length three), we can produce a recommendation based on similarity scores that reflect both the link structure and the trust values on the edges.

Abstract Computational trust has been developed as a novel means of coping with uncertainty within collaborative communities of interacting peers. The idea now offers enourmous potential for use in pervasive mobile environments; however, to date there is little agreement about what computational trust itself means, and what the limitations that emerge from its use are. In this work, we project the idea of computational trust into machine learning terms, showing that trust is a metaphor that helps system designers reason about and exploit the intended deployment scenario to achieve their goals. Viewing a trust model as a strategy to confront a learning problem thus allows us to explore the effect that constraints, such as mobility and user participation, will have on the quantity of information available to learn from; in this work, we demonstrate this idea with a set of experiments on the Reality Mining Dataset. The results highlight that the most successful trust models will be based on strong contextual information about the environment they are to be deployed in. 1

Opportunistic networks enables mobile users to participate in social interactions through applications such as content distribution, flea-market, micro-blogs and round based games. To interact securely, the establishment of trust is vital in a distributed environment. Trust is required to validate an identity and avoid sybil users, select trustworthy interaction partners or to collect useful opinions in a recommender system. Different forms of trust, either based on a social connection (friend), frequent encounter (familiar) or similar interests, can be harnessed in order to best suit the different requirements. Algorithms are proposed in order to evaluate different forms of trust in a distributed manner and combine them for different requirements. Complexity, trust propagation characteristics and security issues are discussed and thoroughly analyzed using synthetic models and real world mobility traces. 1.

Abstract not found