Machine Learning Can Filter False Information and Trolling in Social Media Platforms; Saving America

By Eva Magno / Nov 28, 2016 02:42 PM EST
(Photo : Photo by Justin Sullivan/Getty Images) SAN FRANCISCO - SEPTEMBER 14: Twitter CEO Evan Williams is seen silhouetted against a screen as he shows off the newly revamped Twitter website on September 14, 2010 at Twitter headquarters in San Francisco, California. Twitter launched a new version of the popular social media site in hopes it will be more user friendly. .

With all the trolls, foreign fake news makers, gossip mongers, and media illiterate people acting like journalists, social media platforms can be filled with false information, which leads to hate, ignorance, and even the victory of an unlikely US presidential candidate. One of the best solutions to this, suggested by experts, is to develop efficient filters, such as neural network and deep learning.

Machine learning has been suggested by experts to filter these false information that has been flooding social media platforms, or the internet for that matter. Deep learning can bring high level of accuracy that can solve these very difficult problems, Tech Crunch reported.

This natural-language processing technology can even solve problems in image and voice recognition. It can provide benefits brought by text classification, through algorithms like Doc2vec or Word2vec.

This means words or documents can be turned to vectors, which represent them in numbers. This is called neural embeddings. Algorithms and skilful coding can achieve this.

This is how Google can accurately filter spam to prevent its users from receiving messages from Nigerian princes asking them to hand them their money for a fortune that he will transfer to their bank accounts.

Algorithms that can check facts and judge news stories are available in organizations like Politifact, Media Matters, and Snopes.

By using this technology to detect trolls, users can be protected from trolling by muting them or putting a warning on their posts. This could even filter threats of violence.

Facebook already has its own technology to filter fake news, but it got very controversial during the US elections a few weeks ago, Business Insider reported.

Elad Gil, who used to be an employee in Google and Twitter and is now an advisor for tech companies like AirBnB and Pinterest, criticized Facebook for saying that it is difficult to distinguish fake news from real news.

He said this could mean that Facebook is really bad at machine learning or it simply doesn't want to help solve the issue of false information spreading across social media platforms.

Gil, who also founded biotech company Color Genomics, said he cannot accept Facebook head Mark Zuckerberg's statement saying that sorting out fake news from the real ones needs a complex solution. This is unacceptable, especially with all the technology that Facebook has, according to Gil.