STAGE: Stereotypical Trust Assessment through Graph Extraction

Murat Sensoy, Burcu Yilmaz, Timothy J Norman

Research output: Contribution to journalArticle

8 Citations (Scopus)
25 Downloads (Pure)

Abstract

Bootstrapping trust assessment where there is little or no evidence regarding a subject is a significant challenge for existing trust and reputation systems. When direct or indirect evidence is absent, existing approaches usually assume that all agents are equally trustworthy. This naive assumption makes existing approaches vulnerable to attacks such as Sybil and whitewashing. Inspired by real-life scenarios, we argue that malicious agents may share some common patterns or complex features in their descriptions. If such patterns or features can be detected, they can be exploited to bootstrap trust assessments. Based on this idea, we propose the use of frequent subgraph mining and state-of-the-art knowledge representation formalisms to estimate a priori trust for agents. Our approach first discovers significant patterns that may be used to characterize trustworthy and untrustworthy agents. Then, these patterns are used as features to train a regression model to estimate the trustworthiness of agents. Last, a priori trust for unknown agents (e.g., newcomers) is estimated using the discovered features based on the trained model. Through empirical evaluation, we show that the proposed approach significantly outperforms well-known trust approaches if trustworthiness of agents is correlated with patterns in their descriptions or social networks. Furthermore, we show that the proposed approach performs at least as good as the existing approaches if such correlations do not exist.
Original languageEnglish
Pages (from-to)72-101
Number of pages30
JournalComputational Intelligence
Volume32
Issue number1
Early online date3 Jul 2014
DOIs
Publication statusPublished - Feb 2016

Fingerprint

Graph in graph theory
Trustworthiness
Reputation System
Bootstrapping
Knowledge representation
Knowledge Representation
Estimate
Bootstrap
Social Networks
Subgraph
Mining
Regression Model
Attack
Unknown
Scenarios
Evaluation
Evidence
Model

Keywords

  • trust and reputation
  • Semantic Web
  • graph mining

Cite this

STAGE: Stereotypical Trust Assessment through Graph Extraction. / Sensoy, Murat; Yilmaz, Burcu; Norman, Timothy J.

In: Computational Intelligence, Vol. 32, No. 1, 02.2016, p. 72-101.

Research output: Contribution to journalArticle

Sensoy, Murat ; Yilmaz, Burcu ; Norman, Timothy J. / STAGE: Stereotypical Trust Assessment through Graph Extraction. In: Computational Intelligence. 2016 ; Vol. 32, No. 1. pp. 72-101.
@article{e7e545fd3b6c4419b4ae98d6c1809e51,
title = "STAGE: Stereotypical Trust Assessment through Graph Extraction",
abstract = "Bootstrapping trust assessment where there is little or no evidence regarding a subject is a significant challenge for existing trust and reputation systems. When direct or indirect evidence is absent, existing approaches usually assume that all agents are equally trustworthy. This naive assumption makes existing approaches vulnerable to attacks such as Sybil and whitewashing. Inspired by real-life scenarios, we argue that malicious agents may share some common patterns or complex features in their descriptions. If such patterns or features can be detected, they can be exploited to bootstrap trust assessments. Based on this idea, we propose the use of frequent subgraph mining and state-of-the-art knowledge representation formalisms to estimate a priori trust for agents. Our approach first discovers significant patterns that may be used to characterize trustworthy and untrustworthy agents. Then, these patterns are used as features to train a regression model to estimate the trustworthiness of agents. Last, a priori trust for unknown agents (e.g., newcomers) is estimated using the discovered features based on the trained model. Through empirical evaluation, we show that the proposed approach significantly outperforms well-known trust approaches if trustworthiness of agents is correlated with patterns in their descriptions or social networks. Furthermore, we show that the proposed approach performs at least as good as the existing approaches if such correlations do not exist.",
keywords = "trust and reputation, Semantic Web , graph mining",
author = "Murat Sensoy and Burcu Yilmaz and Norman, {Timothy J}",
year = "2016",
month = "2",
doi = "10.1111/coin.12046",
language = "English",
volume = "32",
pages = "72--101",
journal = "Computational Intelligence",
issn = "0824-7935",
publisher = "Wiley-Blackwell",
number = "1",

}

TY - JOUR

T1 - STAGE: Stereotypical Trust Assessment through Graph Extraction

AU - Sensoy, Murat

AU - Yilmaz, Burcu

AU - Norman, Timothy J

PY - 2016/2

Y1 - 2016/2

N2 - Bootstrapping trust assessment where there is little or no evidence regarding a subject is a significant challenge for existing trust and reputation systems. When direct or indirect evidence is absent, existing approaches usually assume that all agents are equally trustworthy. This naive assumption makes existing approaches vulnerable to attacks such as Sybil and whitewashing. Inspired by real-life scenarios, we argue that malicious agents may share some common patterns or complex features in their descriptions. If such patterns or features can be detected, they can be exploited to bootstrap trust assessments. Based on this idea, we propose the use of frequent subgraph mining and state-of-the-art knowledge representation formalisms to estimate a priori trust for agents. Our approach first discovers significant patterns that may be used to characterize trustworthy and untrustworthy agents. Then, these patterns are used as features to train a regression model to estimate the trustworthiness of agents. Last, a priori trust for unknown agents (e.g., newcomers) is estimated using the discovered features based on the trained model. Through empirical evaluation, we show that the proposed approach significantly outperforms well-known trust approaches if trustworthiness of agents is correlated with patterns in their descriptions or social networks. Furthermore, we show that the proposed approach performs at least as good as the existing approaches if such correlations do not exist.

AB - Bootstrapping trust assessment where there is little or no evidence regarding a subject is a significant challenge for existing trust and reputation systems. When direct or indirect evidence is absent, existing approaches usually assume that all agents are equally trustworthy. This naive assumption makes existing approaches vulnerable to attacks such as Sybil and whitewashing. Inspired by real-life scenarios, we argue that malicious agents may share some common patterns or complex features in their descriptions. If such patterns or features can be detected, they can be exploited to bootstrap trust assessments. Based on this idea, we propose the use of frequent subgraph mining and state-of-the-art knowledge representation formalisms to estimate a priori trust for agents. Our approach first discovers significant patterns that may be used to characterize trustworthy and untrustworthy agents. Then, these patterns are used as features to train a regression model to estimate the trustworthiness of agents. Last, a priori trust for unknown agents (e.g., newcomers) is estimated using the discovered features based on the trained model. Through empirical evaluation, we show that the proposed approach significantly outperforms well-known trust approaches if trustworthiness of agents is correlated with patterns in their descriptions or social networks. Furthermore, we show that the proposed approach performs at least as good as the existing approaches if such correlations do not exist.

KW - trust and reputation

KW - Semantic Web

KW - graph mining

U2 - 10.1111/coin.12046

DO - 10.1111/coin.12046

M3 - Article

VL - 32

SP - 72

EP - 101

JO - Computational Intelligence

JF - Computational Intelligence

SN - 0824-7935

IS - 1

ER -