On February 23, researchers from the Center for Applied Mathematics for Industry (CeMEAI) of the University of Sao Paulo (USP) in Brazil announced a platform to identify “fake news” (fake news). Accuracy up to 96%.
The Vietnam News Agency correspondent in South America quoted CeMEAI experts as saying that the platform applies statistical methods to evaluate text characteristics, such as commonly used vocabulary and grammar patterns. most frequent. Based on an artificial intelligence model, this feature is capable of distinguishing language, vocabulary, and semantic patterns between “fake news” and “true news”, allowing inferences, determining whether a text is fake or not. must be “fake news” or not.
The CeMEAI study used databases collected by researchers from USP, which included 100,000 false and standard news items written in Portuguese published over the past five years.
Chief technology officer at CeMEAI and project coordinator, Francisco Louzada Neto, describes the fight against “fake news” as a race “between cats and mice”. Although “fake news” detection platforms have improved significantly, tools for producing “fake news” are also appearing more and more.
This expert emphasized, the newly launched platform needs to update the database and provide more input information to improve accuracy and increase the ability to identify “fake news”.
According to Brazilian socio-political researchers, the platform is a useful tool in preventing “fake news” related to the COVID-19 pandemic as well as the presidential election in the South American country. this next October.