Fake news is increasingly becoming a menace to our society, on scales inconceivable such as swaying the US presidential elections in 2016. The nature of online news publication has changed, thus traditional fact checking and vetting from potential deception is impossible against the flood arising from content creators, in various formats and genres.
Fake news detection is defined as the prediction of the chances of a particular news article being intentionally deceptive. As biased news is intentionally written to mislead and persuade readers into believing false information, detection becomes all the more difficult. There have been several psychological and cognitive theories that try to explain the natural inability of human beings to differentiate between real and fake news and the influential power of fake news. Two major reasons behind the lack of ability falls on Naïve Realism and Confirmation Bias
From a linguistic perspective, this human inability offers an opportunity to learn and identify patterns and semantic relationships within the content. Deep neural networks have been extensively used in natural language processing tasks such as parsing, language modelling and sentiment analysis, making it apt for this problem.
The fake news detection algorithm is being developed by incorporating algorithms using latest research done in the area of association rule mining, graph databases, and block chain based validation techniques. The primary output would be an API where you can send a news article and the API will give a score for its authenticity based on the in-house algorithm. Webhose data aggregation and other linguistic techniques can be used to provide an authenticity score.
Probyto is collaborating with academia, businesses and organizations to develop a prototype for Fake News. We welcome your feedbacks and happy to partner with you or your organization. Please contact us to get more details.