A Multi-way Parallel Named Entity Annotated Corpus for English, Tamil and Sinhala

TL;DR

This paper introduces a multilingual parallel corpus for English, Tamil, and Sinhala with annotated named entities, establishing new NER benchmarks for low-resource languages and demonstrating applications in NMT.

Contribution

The creation of a multi-way parallel annotated corpus and the evaluation of multilingual models for NER in low-resource languages.

Findings

New benchmark NER results for Sinhala and Tamil.

Analysis of different multilingual language models' NER capabilities.

Demonstrated utility in low-resource neural machine translation.

Abstract

This paper presents a multi-way parallel English-Tamil-Sinhala corpus annotated with Named Entities (NEs), where Sinhala and Tamil are low-resource languages. Using pre-trained multilingual Language Models (mLMs), we establish new benchmark Named Entity Recognition (NER) results on this dataset for Sinhala and Tamil. We also carry out a detailed investigation on the NER capabilities of different types of mLMs. Finally, we demonstrate the utility of our NER system on a low-resource Neural Machine Translation (NMT) task. Our dataset is publicly released: https://github.com/suralk/multiNER.