Semisupervised Neural Proto-Language Reconstruction

TL;DR

This paper introduces a semisupervised neural model for reconstructing proto-languages using limited labeled data and abundant unlabeled data, outperforming existing baselines in the task.

Contribution

It presents a novel neural architecture that leverages linguistic principles and unlabeled data for improved historical language reconstruction.

Findings

Outperforms strong semisupervised baselines

Leverages unlabeled cognate sets effectively

Demonstrates the importance of deterministic transformability

Abstract

Existing work implementing comparative reconstruction of ancestral languages (proto-languages) has usually required full supervision. However, historical reconstruction models are only of practical value if they can be trained with a limited amount of labeled data. We propose a semisupervised historical reconstruction task in which the model is trained on only a small amount of labeled data (cognate sets with proto-forms) and a large amount of unlabeled data (cognate sets without proto-forms). We propose a neural architecture for comparative reconstruction (DPD-BiReconstructor) incorporating an essential insight from linguists' comparative method: that reconstructed words should not only be reconstructable from their daughter words, but also deterministically transformable back into their daughter words. We show that this architecture is able to leverage unlabeled cognate sets to outperform strong semisupervised baselines on this novel task.