Paper Accepted in PACLIC 37

We are happy to announce that one of our papers titled “EDAL: Entropy based Dynamic Attention Loss for HateSpeech Classification”, has been accepted in the “37th Pacific Asia Conference on Language, Information and Computation”. 

In this work, the authors introduce the “Entropy-based Dynamic Attention Loss” (EDAL) to enhance model interpretability by incorporating an additional attention layer. EDAL encourages attention scores that provide valuable insights and boosts the performance of pretrained models during fine-tuning for downstream tasks. We conduct extensive experiments on six diverse datasets, confirming that EDAL effectively enhances classification performance while maintaining interpretability. Additionally, experiments with various pretrained models demonstrate EDAL’s significant performance improvements during fine-tuning. In summary, EDAL holds promise for creating more transparent and reliable hate speech classifiers, contributing to a safer online environment.