Computing Trust as a Form of Presumptive Reasoning

2012

Previous work has used theories of evidence to incorporate belief into trust and reputation systems. Some important questions that remain, however, are how agents might recover reputation lost in disputed transactions, and how new agents with little or no past transaction history might enter the trust network. We attempt to address these issues by extending previous work using the Dempster-Shafer theory of evidence to include formal argumentation. Reasons for past bad transaction assignment can be taken into account in new transactions and discounted by importance. New agents can participate in trust networks by forwarding evidence as arguments in a distributed reputation system. We present our preliminary model on incorporation of argumentation frameworks into trust management systems to support more complex reasoning mechanisms.

Journal of Artificial Intelligence Research, 2011

Leading agent-based trust models address two important needs. First, they show how an agent may estimate the trustworthiness of another agent based on prior interactions. Second, they show how agents may share their knowledge in order to cooperatively assess the trustworthiness of others. However, in real-life settings, information relevant to trust is usually obtained piecemeal, not all at once. Unfortunately, the problem of maintaining trust has drawn little attention. Existing approaches handle trust updates in a heuristic, not a principled, manner. This paper builds on a formal model that considers probability and certainty as two dimensions of trust. It proposes a mechanism using which an agent can update the amount of trust it places in other agents on an ongoing basis. This paper shows via simulation that the proposed approach (a) provides accurate estimates of the trustworthiness of agents that change behavior frequently; and (b) captures the dynamic behavior of the agents. This paper includes an evaluation based on a real dataset drawn from Amazon Marketplace, a leading e-commerce site.

2011

Trust is a concept that has been used in computing to support better decision making. For example, trust can be used in access control. Trust can also be used to support service selection. Although certain elements of trust such as reputation has gained widespread acceptance, a general model of trust has so far not seen widespread usage. This is due to the challenges of implementing a general trust model. In this thesis, a middleware based approach is proposed to address the implementation challenges. The thesis proposes a general trust model known as computational trust. Computational trust is based on research in social psychology. An individual's computational trust is formed with the support of the proposed computational trust architecture. The architecture consists of a middleware and middleware clients. The middleware can be viewed as a representation of the individual that shares its knowledge with all the middleware clients. Each application uses its own middleware client to form computational trust for its decision making needs. Computational trust formation can be adapted to changing circumstances. The thesis also proposed algorithms for computational trust formation. Experiments, evaluations and scenarios are also presented to demonstrate the feasibility of the middleware based approach to computational trust formation.

2008 Third International Conference on Dependability of Computer Systems DepCoS-RELCOMEX, 2008

Trustworthiness of IT systems can be justified using the concept of a trust case. A trust case is an argument structure which encompasses justification and evidence supporting claimed properties of a system. It represents explicitly an expert's way of assessing that a certain object has certain properties. Trust cases can be developed collaboratively on the basis of evidence and justification of varying quality. They can be complex structures impossible to comprehend fully by a non-expert. A postulated model of communicating trust case contents to an 'ordinary' user is an expert acting on user's behalf and communicating his/her assessment to the user. Therefore, a mechanism for issuing and aggregating experts' assessments is required. The paper proposes such a mechanism which enables assessors to appraise strength of arguments included in a trust case. The mechanism uses Dempster-Shafer's model of beliefs to deal with uncertainty resulting from the lack of knowledge of the expert. Different types of argumentation strategies were identified and for each of them appropriate combination rules were presented.

2012

Abstract Trust is a mechanism for managing the uncertainty about autonomous entities and the information they store, and so can play an important role in any decentralized system. As a result, trust has been widely studied in multi-agent systems and related fields such as the semantic web. Here, we introduce a formal system of argumentation that can be used to reason using information about trust.

2011

This paper describes some of our recent work on using argumentation to handle information about trust. We first discuss the importance of trust in computer science in general and in multi-agent systems in particular. We then describe the setting of our work, situating it within the broad area of work on trust.

2012

Agents in open multi-agent systems must deal with the difficult problem of selecting interaction partners in the face of uncertainty about their behaviour. This is especially problematic if they have to interact with an agent they have not interacted with before. In this case they can turn to their peers for information about this potential partner. However, in scenarios where agents may be evaluated according to many different criteria for many different purposes, their peers' evaluations may be mismatched with regards to their own expectations. In this paper we present a novel method, using an argumentation framework, that allows agents to discuss and adapt their trust model. This allows agents to provide, and receive, personalized trust evaluations, better suited to the agent in need, as is shown in a prototypical experiment.

2008

In an open Multi-Agent System, the goals of agents acting on behalf of their owners often conflict with each other. Therefore, a personal agent protecting the interest of a single user cannot always rely on them. Consequently, such a personal agent needs to be able to reason about trusting (information or services provided by) other agents. Existing algorithms that perform such reasoning mainly focus on the immediate utility of a trusting decision, but do not provide an explanation of their actions to the user.

2008

In open settings, the participants are autonomous and there is no central authority to ensure the felicity of their interactions. When agents interact in such settings, each relies upon being able to model the trustworthiness of the agents with whom it interacts. Fundamentally, such models must consider the past behavior of the other parties in order to predict their future behavior. Further, it is sensible for the agents to share information via referrals to trustworthy agents. Much progress has recently been made on probabilistic trust models including those that support the aggregation of information from multiple sources. However, current models do not support trust updates, leaving updates to be handled in an ad hoc manner.

2013

Trust relationships occur naturally in many diverse contexts such as collaborative systems, e-commerce, interpersonal interactions, social networks, and semantic sensor web. As agents providing content and services become increasingly removed from the agents that consume them, the issue of robust trust inference and update becomes critical. There is a need to find online substitutes for traditional (direct or face-to-face) cues to derive measures of trust, and create efficient and robust systems for managing trust in order to support decision-making. Unfortunately, there is neither a universal notion of trust that is applicable to all domains nor a clear explication of its semantics or computation in many situations. We motivate the trust problem, explain the relevant concepts, summarize research in modeling trust and gleaning trustworthiness, and discuss challenges confronting us. The goal is to provide a comprehensive broad overview of the trust landscape, with the nitty-gritties of a handful of approaches. We also provide details of the theoretical underpinnings and comparative analysis of Bayesian approaches to binary and multi-level trust, to automatically determine trustworthiness in a variety of reputation systems including those used in sensor networks, e-commerce, and collaborative environments. Ultimately, we need to develop expressive trust networks that can be assigned objective semantics.