Tackling the Low-resource Challenge for Canonical Segmentation

TL;DR

This paper investigates methods for canonical morphological segmentation in low-resource languages, introducing new models that outperform existing ones in simulated settings but still face challenges with truly low-resource languages.

Contribution

It proposes two novel models for low-resource morphological segmentation and evaluates their performance on both high-resource and truly low-resource languages.

Findings

New models outperform existing ones by up to 11.4% accuracy in low-resource settings.

Accuracy remains over 50% in simulated low-resource scenarios but drops significantly for truly low-resource languages.

Canonical segmentation remains a challenging task for languages with very limited data.

Abstract

Canonical morphological segmentation consists of dividing words into their standardized morphemes. Here, we are interested in approaches for the task when training data is limited. We compare model performance in a simulated low-resource setting for the high-resource languages German, English, and Indonesian to experiments on new datasets for the truly low-resource languages Popoluca and Tepehua. We explore two new models for the task, borrowing from the closely related area of morphological generation: an LSTM pointer-generator and a sequence-to-sequence model with hard monotonic attention trained with imitation learning. We find that, in the low-resource setting, the novel approaches outperform existing ones on all languages by up to 11.4% accuracy. However, while accuracy in emulated low-resource scenarios is over 50% for all languages, for the truly low-resource languages Popoluca and Tepehua, our best model only obtains 37.4% and 28.4% accuracy, respectively. Thus, we conclude that canonical segmentation is still a challenging task for low-resource languages.