DeFacto—Temporal and multilingual Deep Fact Validation

Abstract

One of the main tasks when creating and maintaining knowledge bases is to validate facts and provide sources for them in order to ensure correctness and traceability of the provided knowledge. So far, this task is often addressed by human curators in a three-step process: issuing appropriate keyword queries for the statement to check using standard search engines, retrieving potentially relevant documents and screening those documents for relevant content. The drawbacks of this process are manifold. Most importantly, it is very time-consuming as the experts have to carry out several search processes and must often read several documents. In this article, we present DeFacto (Deep Fact Validation) -an algorithm able to validate facts by finding trustworthy sources for them on the Web. DeFacto aims to provide an effective way of validating facts by supplying the user with relevant excerpts of web pages as well as useful additional information including a score for the confidence DeFacto has in the correctness of the input fact. To achieve this goal, DeFacto collects and combines evidence from web pages written in several languages. In addition, DeFacto provides support for facts with a temporal scope, i.e., it can estimate in which time frame a fact was valid. Given that the automatic evaluation of facts has not been paid much attention to so far, generic benchmarks for evaluating these frameworks were not previously available. We thus also present a generic evaluation framework for fact checking and make it publicly available.

Advances in Machine Learning for the Behavioral Sciences

“…One of the main uses of Knowledge graphs in the behavioral sciences is in the study of the spread of disinformation (Luca Ciampaglia et al, 2015;Fernandez and Alani, 2018). DBpedia is used for computational fact-checking in several systems, including DeFacto (Gerber et al, 2015). Knowledge graphs are also used to enhance understanding of the text by linking keywords and entities appearing in text to more general concepts.…”

Measuring Accuracy of Triples in Knowledge Graphs

“…These processes could be very noisy, and the created knowledge graphs are unlikely to be fully correct. There is an increasing interest in quality assessment for knowledge graphs [1] [4] [17] [20] [8] [6]. Some approaches focus on completing or correcting entity type information, while others target towards relations between entities, or interlinks between di↵erent knowledge graphs.…”

“…A confidence score is assigned to a source triple that is validated to represent the accuracy of this triple. Our approach searches external information for assessing the correctness of triples, which is similar to [8] [6]. The main di↵erence is that we explore other semantic web knowledge bases to find evidence while [6] searches proofs from the Web for validating facts.…”

“…There is a growing body of work on fact validation [6] [20] [10] [5]. This literature can be categorised into two groups in terms of the sources utilised: (1) approaches such as [20] and [10] using internal information (i.e.…”

“…This literature can be categorised into two groups in terms of the sources utilised: (1) approaches such as [20] and [10] using internal information (i.e. the knowledge graph itself) for proofs; and (2) approaches such as [6] exploring external information as sources for evidences. Our approach is similar to methods of the second kind, which validate triples using external sources.…”