Work

Domain Classification-based Source-specific Term Penalization for Domain Adaptation in Hate-speech Detection

Year: 2022

Type: preprint

Abstract: State-of-the-art approaches for hate-speech detection usually exhibit poor performance in out-of-domain settings. This occurs, typically, due to classifiers overemphasizing source-specific information that negatively impacts its domain invariance. Prior work has attempted to penalize terms related to hate-speech from manually curated lists using feature attribution methods, which quantify the importance assigned to input terms by the classifier when making a prediction. We, instead, propose a domain adaptation approach that automatically extracts and penalizes source-specific terms using a domain classifier, which learns to differentiate between domains, and feature-attribution scores for hate-speech classes, yielding consistent improvements in cross-domain evaluation.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2209.08681

Authors: Bose, Tulika, Aletras, Nikolaos, Illina, Irina, Fohr, Dominique

Institutions: Centre National de la Recherche Scientifique, Institut national de recherche en sciences et technologies du numérique, Laboratoire Lorrain de Recherche en Informatique et ses Applications, Université de Lorraine, University of Sheffield

Language: English

Topic: Hate Speech and Cyberbullying Detection

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

Funders: Agence Nationale de la Recherche, Centre National de la Recherche Scientifique

Awards: G1166465053, G1266851076, G6554659165

OpenAlex ID: W4296544267

Cites:

Cited by: 3

Related to: 10