Neural Multi-Source Morphological Reinflection

TL;DR

This paper introduces a multi-source morphological reinflection task, proposes a novel multi-encoder neural architecture, and demonstrates its superior performance over single-source models, along with releasing a new dataset for future research.

Contribution

It presents a new multi-source reinflection task, a novel multi-encoder neural model, and a dataset to advance research in morphological reinflection.

Findings

Multi-source models outperform single-source models.

The proposed architecture effectively leverages multiple source forms.

A new dataset for multi-source morphological reinflection is published.

Abstract

We explore the task of multi-source morphological reinflection, which generalizes the standard, single-source version. The input consists of (i) a target tag and (ii) multiple pairs of source form and source tag for a lemma. The motivation is that it is beneficial to have access to more than one source form since different source forms can provide complementary information, e.g., different stems. We further present a novel extension to the encoder- decoder recurrent neural architecture, consisting of multiple encoders, to better solve the task. We show that our new architecture outperforms single-source reinflection models and publish our dataset for multi-source morphological reinflection to facilitate future research.