What do you believe is the single most important factor driving up the cost of living in Nigeria?

Researchers have developed a new way to automatically detect hate speech on social media platforms more accurately and consistently using a new multi-task learning (MTL) model; a type of machine learning model that works across multiple datasets.

Associate Professor Marian-Andrei Rizoiu, Head of the Behavioural Data Science Lab at the University of Technology Sydney (UTS) is working on the frontline in the fight against online misinformation and hate speech. His interdisciplinary research combines computer and social sciences, to better understand and predict human attention in the online environment, including the types of speech that influence and polarize opinion on digital channels.

"As social media becomes a significant part of our daily lives, automatic identification of hateful and abusive content is vital in combating the spread of harmful content and preventing its damaging effects," said Associate Professor Rizoiu.

"Designing effective automatic detection of hate speech is a significant challenge. Current models are not very effective in identifying all the different types of hate speech, including racism, sexism, harassment, incitement to violence and extremism.

• Read More