Indiana University (IU) researchers have developed a computational method that can exploit any body of knowledge to help human fact-checkers.
The computational fact-checker assigns truth scores to statements concerning history, geography, and entertainment, as well as random statements taken from Wikipedia. In several experiments, the system consistently matched the assessment of human fact-checkers in terms of their certitude about the accuracy of the statements.
"Our experiments point to methods to abstract the vital and complex human task of fact-checking into a network analysis problem, which is easy to solve computationally," says Giovanni Luca Ciampaglia, a postdoctoral fellow at IU's Center for Complex Networks and Systems Research.
The researchers used factual information from Wikipedia to build a knowledge graph with 3 million concepts and 23 million links between them. They then applied the algorithm to answer simple questions related to geography, history, and entertainment. The researchers also tested the algorithm against human fact-checkers, and found a positive correlation between the truth scores produced by the algorithm and the answers provided by humans.
"With increasing reliance on the Internet as a source of information, we need tools to deal with the misinformation that reaches us every day," says Filippo Menczer, director of IU's Center for Complex Networks and Systems Research.
From IU Bloomington Newsroom
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found