Minimal Supervision for Morphological Inflection

TL;DR

This paper introduces a minimal supervision approach for morphological inflection that bootstraps labeled data from as few as five examples using unlabeled text, leveraging regularities in morphology to improve accuracy across languages.

Contribution

The work presents a two-phased method exploiting orthographic and semantic regularities to reduce the need for extensive labeled data in morphological inflection tasks.

Findings

Orthographic regularities enable effective inflection in simple languages.

Combining orthographic and semantic cues improves performance on complex morpho-phonological systems.

Few labeled examples suffice for high accuracy in many cases.

Abstract

Neural models for the various flavours of morphological inflection tasks have proven to be extremely accurate given ample labeled data -- data that may be slow and costly to obtain. In this work we aim to overcome this annotation bottleneck by bootstrapping labeled data from a seed as little as {\em five} labeled paradigms, accompanied by a large bulk of unlabeled text. Our approach exploits different kinds of regularities in morphological systems in a two-phased setup, where word tagging based on {\em analogies} is followed by word pairing based on {\em distances}. We experiment with the Paradigm Cell Filling Problem over eight typologically different languages, and find that, in languages with relatively simple morphology, orthographic regularities on their own allow inflection models to achieve respectable accuracy. Combined orthographic and semantic regularities alleviate difficulties with particularly complex morpho-phonological systems. Our results suggest that hand-crafting many tagged examples might be an unnecessary effort. However, more work is needed in order to address rarely used forms.