AI for #MeToo: Training algorithms to spot online trolls
Source: California Institute of Technology
Researchers at Caltech have demonstrated that machine-learning algorithms can monitor online social media conversations as they evolve, which could one day lead to an effective and automated way to spot online trolling.
The project unites the labs of artificial intelligence (AI) researcher Anima Anandkumar, Bren Professor of Computing and Mathematical Sciences, and Michael Alvarez, professor of political science. Their work was presented on December 14 at the AI for Social Good workshop at the 2019 Conference on Neural Information Processing Systems in Vancouver, Canada. Their research team includes Anqi Liu, postdoctoral scholar; Maya Srikanth, a junior at Caltech; and Nicholas Adams-Cohen (MS '16, Ph.D. '19) of Stanford University.
"This is one of the things I love about Caltech: the ability to bridge boundaries, developing synergies between social science and, in this case, computer science," Alvarez says.
Prevention of online harassment requires rapid detection of offensive, harassing, and negative social media posts, which in turn requires monitoring online interactions. Current methods to obtain such social media data are either fully automated and not interpretable or rely on a static set of keywords, which can quickly become outdated. Neither method is very effective, according to Srikanth.
"It isn't scalable to have humans try to