The method, dubbed the Multi-Modal Discussion Transformer (mDT), can understand the relationship between text and images as well as put comments in greater context, unlike previous hate speech detection methods.
"We believe that by taking a community-centred approach in our applications of AI, we can help create safer online spaces for all." - Liam Hebert, a Waterloo computer science PhD student and the first author of the study.
"Context is very important when understanding hate speech. For example, the comment 'That's gross!' might be innocuous by itself, but its meaning changes dramatically if it's in response to a photo of pizza with pineapple versus a person from a marginalized group." - Liam Hebert.
#hate-speech-detection #machine-learning #social-media-analysis #research-organizations #university-of-waterloo
[
Collection
]
[
|
...
]