Does Typological Blinding Impede Cross-Lingual Sharing?

TL;DR

This paper investigates whether explicitly using typological features enhances cross-lingual models, finding that blinding models to such features significantly reduces performance, while encouraging typological sharing can improve results.

Contribution

It demonstrates that explicit typological information is crucial for cross-lingual sharing, challenging prior assumptions of its limited utility.

Findings

Blinding models to typology reduces performance.

Encouraging typological sharing improves cross-lingual transfer.

Typological cues are implicitly learned by models in cross-lingual settings.

Abstract

Bridging the performance gap between high- and low-resource languages has been the focus of much previous work. Typological features from databases such as the World Atlas of Language Structures (WALS) are a prime candidate for this, as such data exists even for very low-resource languages. However, previous work has only found minor benefits from using typological information. Our hypothesis is that a model trained in a cross-lingual setting will pick 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. Our model is based on a cross-lingual architecture in which the latent weights governing the sharing between languages is learnt during training. We show that (i) preventing this model from exploiting typology severely reduces performance, while a control experiment reaffirms that (ii) encouraging sharing according to typology somewhat improves performance.