Abstract

I will discuss the usefulness of typological features in NLP , particularly in cross-lingual settings. On the one hand, typological features from databases such as the World Atlas of Language Structures (WALS) seem promising for cross-lingual NLP , as annotations for useful aspects of language exist even for very low-resource languages. Furthermore, missing features in WALS can be predicted with relatively high success, and has been the focus of much recent work (e.g. in the SIGTYP 2020 shared task). When it comes to application of these features, however, previous work has only found minor benefits from using typological information in actual NLP modelling. In recent work (to appear at EACL 2021 ), we hypothesised that these minor gains might stem from that a model trained in a cross-lingual setting picks up on typological cues from the input data, thus overshadowing the utility of explicitly using such features. We verify this hypothesis by blinding a model to typological information, and investigate how cross-lingual sharing and performance is impacted. While this sheds some light on the matter, I want to further explore the question of the usefulness of typological feature prediction in general, and the use of such features in NLP .