Company Meta developed Sphere artificial intelligence algorithm that automatically checks Wikipedia footnotes. He can also suggest alternatives if a passage is found to be inaccurate.
(1/2) Today, we’re announcing the first model capable of automatically verifying hundreds of thousands of citations at once. Read more: https://t.co/c0nnvhuQx7 pic.twitter.com/UTSEYUmCRZ
— MetaAI (@MetaAI) July 11, 2022
Sphere uses a natural language comprehension (NLU) transformation model that tries to understand the different relationships of words and phrases in a sentence.
The Sphere database consists of 134 million web pages. The natural language understanding algorithm parses the footnotes in encyclopedia articles and looks for a single source to verify each statement.
The company says Sphere can analyze “tens of thousands” of quotes at once. Meta noted that the electronic encyclopedia is replenished with 17,000 articles at a time, containing many footnotes that are difficult to check manually.
To illustrate the possibilities of AI, Meta showed an example of an incomplete quote found by a model on a Wikipedia page about Native Americans. blackfoot. In the “Famous Persons” section (in the English version of the article), Joe Hipp is mentioned as the first representative of the nation to fight for the WBA world heavyweight title.
However, the linked website could not confirm this fact. Searching the Sphere database, the model found a more relevant quote in a 2015 article in the Great Falls Tribune. The algorithm noted the following passage:
“In 1989, at the end of my career, [Марвин] Camel fought Joe Hipp of the Blackfoot Nation. Hipp, who became the first Native American to challenge a world heavyweight champion, said the fight was one of the strangest of his career.”
The found passage does not directly mention boxing. Sphere found a suitable link thanks to natural language capabilities, the developers emphasized.
The company believes that in the future the tool will help fight disinformation on Facebook.
“More generally, we hope that our work can be used to verify facts and improve the overall validity of information on the Internet,” the creators of the model said.
In addition, Meta plans to create a platform that Wikipedia editors can use to systematically review and correct footnotes.
The source code for the model is available at GitHub.
Recall that in July, Meta introduced an AI translator that supports 200 languages.
In April, the tech giant talked about the development of the open source OPT-175B language model.
Subscribe to Cryplogger news in Telegram: Cryplogger AI – all the news from the world of AI!
Found a mistake in the text? Select it and press CTRL+ENTER
