David DeAngelis

Attitude Driven Team Formation using Multi-Dimensional Trust

When agents form a team to solve a given problem, a critical step in improving performance is selecting beneficial teammates by identifying the helpfulness of other agents. To maximize its performance, an agent must consider the trustworthiness of potential teammates relative to multiple behavioral constraints. This multidimensional trustworthiness assessment is shown to be of significant benefit in solving the team formation problem. This research introduces the concept of attitude to assert how much an agent should trust other agents by identifying the most influential facet among multiple trustworthiness assessments. In this sense, attitudes define how an agent selects beneficial teammates given different situations. In addition, this research shows how those attitudes are learned and aid in teammate selection.

Expertise Modeling and Recommendation in Online Question and Answer Forums

Question and answer forums provide a method of connecting users and resources that can leverage both the static and dynamic (live) capabilities of a network of human users. We present a recommender for selecting the most appropriate responders given a question. The goal of this work is to encourage expert participation in QA forums by reducing the time investment needed by an expert to find a suitable question, decrease responder load, and to increase questioner confidence in the responses of others. The two primary contributions of this work are: 1. a generative model for characterizing the production of content in an online question and answer forum and 2. a decision theoretic framework for recommending expert participants while maintaining questioner satisfaction and distributing responder load. We have also developed two new metrics tailored to QA forums: responder load and questioner satisfaction. These metrics are used to evaluate the performance of our recommender system on datasets harvested from Yahoo! Answers. Experiments across several topic domains demonstrate our system’s ability to predict responder identities and suggest new responders that are likely to have the appropriate expertise.

Agent trust evaluation and team formation in heterogeneous organizations

This demonstration highlights different aspects of the bottom-up assembly of multi-agent teams; illustrating trust evaluation of potential partners via experience- and reputation-based trust models, multi-dimensional trust evaluation of potential partners, task selection through personality-based modeling and team selection strategies that maximize a teams ability to function in dynamic environments. The demonstration format will be a software live demo with supporting slide shows.

Security applications of trust in multi-agent systems

The concept of trust as presented here focuses on the trustworthiness, or reliability, of information and information sources. Decision makers, or agents, can create judgments based on previous experience with other agents and by reputation information received from allied agents. These judgments, or trust assessments, are used to predict the behavior of other agents and analyze the trustworthiness, truthfulness, or quality of information. Research concepts have been developed within the trust community, and they are most commonly applied to multi-agent systems research. This work attempts to show that trust research can be directly applied to security problems. Modern trust concepts enforce soft security which can be applied in addition to conventional security methods to build a more robust secure system. This article examines two trust based techniques and demonstrates their basic effectiveness using empirical experimentation. These techniques are then applied in a case study drawn from a more robust domain concerning confidential message transmission. The benefits of applying trust-based techniques to secure a system are measurable, and the costs associated with such techniques are scalable to even the most resource constrained systems.

Teammate Selection Using Multi-dimensional Trust and Attitude Models

Multi-dimensional trustworthiness assessments have been shown significantly beneficial to agents when selecting appropriate teammates to achieve a given goal. Reliability, quality, availability, and timeliness define the behavioral constraints of the proposed multi-dimensional trust (MDT) model. Given the multi-dimensional trust model in this research, an agent learns to identify the most beneficial teammates given different situations by prioritizing each dimension differently. An agents attitudes towards rewards, risks and urgency are used to drive an agents prioritization of dimensions in a MDT model. Each agent is equipped with a reinforcement learning mechanism with clustering technique to identify its optimal set of attitudes and change its attitudes when the environment changes. Experimental results show that changing attitudes to give preferences for respective dimensions in the MDT, and consequently, teammate selection based on the situation offer a superior means of finding the best teammates for goal achievement.

Systemic Reciprocal Rewards: Motivating Expert Participation in Online Communities with a Novel Class of Incentives

Online communities such as question and answer (QA) systems are growing rapidly and we increasingly rely on them for valuable information and entertainment. However, finding meaningful rewards to motivate participation from the most qualified users, or experts, presents researchers with two main challenges: identifying these users and (2) rewarding their participation. Using an interdisciplinary theoretical framework, we illustrate possibilities for identifying and motivating the most valuable contributors to online communities. We suggest that access to peer-generated content can directly motivate people to apply their own expertise, thereby generating more content. Survey data from 380 participants suggests that users strongly prefer a novel class of incentives—reciprocal systemic rewards—to traditional achievement-based rewards. Overall, this research presents important considerations for many different types of online communities, including social networking and news aggregation sites. Systemic Reciprocal Rewards: Motivating Expert Participation in Online Communities with a Novel Class of Incentives

Modeling Virtual Footprints

Entities interacting on the web establish their identity by creating virtual personas. These entities, or agents, can be human users or software-based. This research models identity using the Entity-Persona Model, a semantically annotated social network inferred from the persistent traces of interaction between personas on the web. A Persona Mapping Algorithm is proposed which compares the local views of personas in their social network referred to as their Virtual Signatures, for structural and semantic similarity. The semantics of the Entity-Persona Model are modeled by a vector space model of the text associated with the personas in the network, which allows comparison of their Virtual Signatures. This enables all the publicly accessible personas of an entity to be identified on the scale of the web. This research enables an agent to identify a single entity using multiple personas on different networks, provided that multiple personas exhibit characteristic behavior. The agent is able to increase the trustworthiness of on-line interactions by establishing the identity of entities operating under multiple personas. Consequently, reputation measures based on on-line interactions with multiple personas can be aggregated and resolved to the true singular identity.