Milimili. Collecting Parallel Data via Crowdsourcing

TL;DR

This paper introduces a crowdsourcing-based method for collecting parallel corpora, providing a cost-effective alternative to professional translation, and shares experimental data for Chechen-Russian and Fula-English language pairs.

Contribution

It presents a novel crowdsourcing methodology for parallel data collection and releases new experimental datasets for under-resourced language pairs.

Findings

Crowdsourcing is a viable, cost-effective approach for collecting parallel corpora.

Experimental datasets for Chechen-Russian and Fula-English are now available.

The quality of crowdsourced data varies but can be useful for NLP tasks.

Abstract

We present a methodology for gathering a parallel corpus through crowdsourcing, which is more cost-effective than hiring professional translators, albeit at the expense of quality. Additionally, we have made available experimental parallel data collected for Chechen-Russian and Fula-English language pairs.