Algorithm Can Root Out Dangerous Internet Trolls

A new algorithm can detect groups of online users who spread fake news, or promote violent behavior and extremism.

It’s been developed by researchers at the Ben-Gurion University of the Negev and it analyzes social media networks to detect malicious communities. 

They applied their method to two real-world networks: on a community within Reddit, a popular American content rating website, and a Hebrew Wikipedia article, whose editors were classified as users for the purpose of the study. The results demonstrated that the method successfully identified malicious users in both cases. 

Their machine learning algorithm can help identify groups of fake profiles spreading fake news, and prevent targeted violence toward minorities by uncovering hatred-inciting communities in an online social network. 

Industry and academic researchers have proposed various solutions over the last two decades – but they mostly fail when the fake users have similar properties as regular users and can ‘fool’ the systems. 

“Our method is generic. Therefore, it can potentially work on different types of social media platforms. We tested it on several different types of networks, such as Reddit and Wikipedia (which is also a type of social network),” said Dr. Michael Fire, one of the researchers who led the study.

“Our method is based solely on network structural properties. That makes our method independent of vertices’ attributes (the connections between users online).

“Thus, it is agnostic to the domain. When comparing our algorithm with other algorithms, it performed better on simulation and real-world data in many cases. It successfully detected groups of anomalous users’ communities who presented peculiar online activity.”

The researchers’ findings were published in the peer-reviewed journal Neural Processing Letters.