ur-iw-hnt at GermEval 2021: An Ensembling Strategy with Multiple BERT Models
This work addresses the problem of content moderation in German social media, but it is incremental as it applies existing ensembling techniques to a new dataset.
The authors tackled the GermEval 2021 task of identifying toxic, engaging, and fact-claiming comments by using an ensembling strategy with multiple BERT models, finding that ensemble models outperformed single models and BERTweet was the best individual performer across all subtasks.
This paper describes our approach (ur-iw-hnt) for the Shared Task of GermEval2021 to identify toxic, engaging, and fact-claiming comments. We submitted three runs using an ensembling strategy by majority (hard) voting with multiple different BERT models of three different types: German-based, Twitter-based, and multilingual models. All ensemble models outperform single models, while BERTweet is the winner of all individual models in every subtask. Twitter-based models perform better than GermanBERT models, and multilingual models perform worse but by a small margin.