4 accepted papers
Continual pre-training has long been considered the default strategy for adapting models to non-English languages, but struggles with initializing new embeddings, particularly for non-Latin scripts. In this work, we propose EnerGIZAr, a novel methodology that improves continual pre-training by lever…
This paper presents preliminary experiments for the lemmatisation of unedited, Byzantine Greek epigrams. This type of Greek is quite different from its classical ancestor, mostly because of its orthographic inconsistencies. Existing lemmatisation algorithms display an accuracy drop of around 30pp wh…
Given the omnipresence of social media in our society, thoughts and opinions are being shared online in an unprecedented manner. This means that both positive and negative emotions can be equally and freely expressed. However, the negativity bias posits that human beings are inherently drawn to and…
In this research, we present pilot experiments to distil monolingual models from a jointly trained model for 102 languages (mBERT). We demonstrate that it is possible for the target language to outperform the original model, even with a basic distillation setup. We evaluate our methodology for 6 lan…