Easy Adaptation to Mitigate Gender Bias in Multilingual Text Classification
This work addresses gender bias in multilingual text classification, which is an incremental improvement over existing monolingual approaches.
The paper tackled gender bias in multilingual text classification by applying a domain adaptation model to treat gender as domains, and demonstrated its effectiveness on hate speech detection and rating prediction tasks compared to three fair-aware baselines.
Existing approaches to mitigate demographic biases evaluate on monolingual data, however, multilingual data has not been examined. In this work, we treat the gender as domains (e.g., male vs. female) and present a standard domain adaptation model to reduce the gender bias and improve performance of text classifiers under multilingual settings. We evaluate our approach on two text classification tasks, hate speech detection and rating prediction, and demonstrate the effectiveness of our approach with three fair-aware baselines.