Does Character-level Information Always Improve DRS-based Semantic Parsing?

Tomoya Kurosawa, Hitomi Yanaka

Even in the era of massive language models, it has been suggested that character-level representations improve the performance of neural models. The state-of-the-art neural semanti…