Abstract : Rapidly changing social media content calls for robust and generalisable abuse detection models. However, the state-of-the-art supervised models display degraded performance when they are evaluated on abusive comments that differ from the training corpus. We investigate if the performance of supervised models for cross-corpora abuse detection can be improved by incorporating additional information from topic models, as the latter can infer the latent topic mixtures from unseen samples. In particular, we combine topical information with representations from a model tuned for classifying abusive comments. Our performance analysis reveals that topic models are able to capture abuse-related topics that can transfer across corpora, and result in improved generalisability.
https://hal.inria.fr/hal-03212196 Contributor : Tulika BoseConnect in order to contact the contributor Submitted on : Thursday, April 29, 2021 - 2:01:07 PM Last modification on : Saturday, February 5, 2022 - 3:08:38 AM Long-term archiving on: : Friday, July 30, 2021 - 6:41:23 PM