[Paper Review] Truth Discovery to Resolve Object Conflicts in Linked Data

Wenqiang Liu|arXiv (Cornell University)|Sep 1, 2015

Semantic Web and Ontologies33 references3 citations

TL;DR

This paper proposes TruthDiscover, a novel truth discovery approach for resolving object conflicts in scale-free Linked Data, where most sources provide few conflicting objects while a few provide many. It leverages topological properties of the Source Belief Graph to estimate prior source reliability and uses a Hidden Markov Random Field to model interdependencies between objects, improving trustworthiness estimation in skewed data distributions. The method significantly outperforms existing approaches on six real-world datasets.

ABSTRACT

In the community of Linked Data, anyone can publish their data as Linked Data on the web because of the openness of the Semantic Web. As such, RDF (Resource Description Framework) triples described the same real-world entity can be obtained from multiple sources; it inevitably results in conflicting objects for a certain predicate of a real-world entity. The objective of this study is to identify one truth from multiple conflicting objects for a certain predicate of a real-world entity. An intuitive principle based on common sense is that an object from a reliable source is trustworthy; thus, a source that provide trustworthy object is reliable. Many truth discovery methods based on this principle have been proposed to estimate source reliability and identify the truth. However, the effectiveness of existing truth discovery methods is significantly affected by the number of objects provided by each source. Therefore, these methods cannot be trivially extended to resolve conflicts in Linked Data with a scale-free property, i.e., most of the sources provide few conflicting objects, whereas only a few sources have many conflicting objects. To address this challenge, we propose a novel approach called TruthDiscover to identify the truth in Linked Data with a scale-free property. Two strategies are adopted in TruthDiscover to reduce the effect of the scale-free property on truth discovery. First, this approach leverages the topological properties of the Source Belief Graph to estimate the priori beliefs of sources, which are utilized to smooth the trustworthiness of sources. Second, this approach utilizes the Hidden Markov Random Field to model the interdependencies between objects to estimate the trust values of objects accurately. Experiments are conducted in the six datasets to evaluate TruthDiscover.

Motivation & Objective

• To address the challenge of truth discovery in Linked Data with a scale-free distribution of conflicting objects across sources.
• To reduce the negative impact of skewed source contribution patterns on trust estimation in truth discovery.
• To improve the accuracy of truth identification by modeling interdependencies between conflicting objects.
• To develop a method robust to the imbalance in source contributions, common in real-world Linked Data.
• To provide a reliable truth discovery framework that adapts to the structural properties of real-world data graphs.

Proposed method

• It constructs a Source Belief Graph to represent source-object relationships and uses its topological properties to derive prior beliefs for sources.
• It applies a prior belief smoothing mechanism based on the graph's structure to improve initial reliability estimates.
• It models interdependencies between conflicting objects using a Hidden Markov Random Field (HMRF).
• The HMRF captures conditional dependencies among objects to refine trust scores beyond individual source reliability.
• It iteratively updates source reliability and object trust values using a joint inference process.
• The method jointly estimates source trustworthiness and object truthfulness through probabilistic inference on the HMRF model.

Experimental results

Research questions

• RQ1 How can truth discovery be made robust to the scale-free distribution of conflicting objects in Linked Data?
• RQ2 To what extent can topological features of the source-object graph improve prior belief estimation for sources?
• RQ3 Can modeling interdependencies between conflicting objects enhance truth discovery accuracy?
• RQ4 How does TruthDiscover perform compared to existing truth discovery methods on real-world datasets?
• RQ5 What is the impact of skewed source contribution patterns on traditional truth discovery approaches?

Key findings

• TruthDiscover significantly improves truth discovery accuracy on six real-world datasets compared to baseline methods.
• The use of topological priors in the Source Belief Graph reduces bias from sparse contributors and enhances reliability estimation.
• The HMRF-based modeling of object interdependencies leads to more accurate trust score propagation across conflicting values.
• The method demonstrates robust performance even when a small number of sources dominate the conflict data.
• Experiments show that TruthDiscover outperforms existing truth discovery techniques, especially in scenarios with high source contribution imbalance.
• The approach effectively mitigates the adverse effects of scale-free data distributions on truth discovery.