Transformer Ensembles for Sexism Detection
This work addresses the problem of detecting sexism in online content for researchers and practitioners, but it is incremental as it applies existing ensemble methods to a specific dataset.
The authors tackled sexism detection in social media by using ensembles of Transformer-based models, achieving an accuracy of 0.767 and F1 score of 0.766 for binary classification, and 0.623 accuracy and 0.535 F1 score for multi-class classification.
This document presents in detail the work done for the sexism detection task at EXIST2021 workshop. Our methodology is built on ensembles of Transformer-based models which are trained on different background and corpora and fine-tuned on the provided dataset from the EXIST2021 workshop. We report accuracy of 0.767 for the binary classification task (task1), and f1 score 0.766, and for the multi-class task (task2) accuracy 0.623 and f1-score 0.535.