White-Box Attacks on Hate-speech BERT Classifiers in German with Explicit and Implicit Character Level Defense
This work addresses security vulnerabilities in hate speech detection systems for German language applications, representing an incremental improvement in adversarial robustness testing.
The researchers evaluated the adversarial robustness of BERT models on German hate speech datasets by developing novel white-box character and word level attacks, and compared two character-level defense strategies to assess their effectiveness.
In this work, we evaluate the adversarial robustness of BERT models trained on German Hate Speech datasets. We also complement our evaluation with two novel white-box character and word level attacks thereby contributing to the range of attacks available. Furthermore, we also perform a comparison of two novel character-level defense strategies and evaluate their robustness with one another.