LORELEI Multiway Translated Text

Item Name: LORELEI Multiway Translated Text
Author(s): Jennifer Tracey, Stephanie Strassel, David Graff, Jonathan Wright, Song Chen, Dana Delgado, Michael Arrigo
LDC Catalog No.: LDC2026T06
ISLRN: 011-662-998-290-3
DOI: https://doi.org/10.35111/fm0k-2a20
Release Date: June 15, 2026
Member Year(s): 2026
DCMI Type(s): Software, Text
Data Source(s): newswire, varied
Project(s): LORELEI
Application(s): cross-language transfer, machine translation
Language(s): Arabic, Amharic, Bengali, Mandarin Chinese, English, Persian, Hausa, Hindi, Hungarian, Indonesian, Russian, Somali, Spanish, Swahili, Tamil, Tagalog, Thai, Turkish, Uzbek, Vietnamese, Gambian Wolof, Yoruba, Zulu, Akan
Language ID(s): ara, amh, ben, cmn, eng, fas, hau, hin, hun, ind, rus, som, spa, swa, tam, tgl, tha, tur, uzb, vie, wof, yor, zul, aka
License(s): LDC User Agreement for Non-Members
Online Documentation: LDC2026T06 Documents
Licensing Instructions: Subscription & Standard Members, and Non-Members
Citation: Tracey, Jennifer, et al. LORELEI Multiway Translated Text LDC2026T06. Web Download. Philadelphia: Linguistic Data Consortium, 2026.
Related Works: View

Introduction

LORELEI Multiway Translated Text consists of a fixed set of English texts (around 100,000 words) translated into 24 languages. It was developed by the Linguistic Data Consortium (LDC) for the DARPA LORELEI Program; the translations were included in the LORELEI representative language packs created by LDC in 2016-2019.

The LORELEI (Low Resource Languages for Emergent Incidents) program was concerned with building human language technology for low resource languages in the context of emergent situations like natural disasters or disease outbreaks. Linguistic resources for LORELEI include Representative Language Packs and Incident Language Packs for over two dozen low resource languages, comprising data, annotations, basic natural language processing tools, lexicons and grammatical resources. Representative languages were selected to provide broad typological coverage, while incident languages were selected to evaluate system performance on a language whose identity was disclosed at the start of the evaluation.

Data

The common set of 100,000 English words translated into each representative language was composed of English news documents (50%), LORELEI-domain English news documents (25%), and a phrasebook and elicitation corpus (25%). The phrasebook contained everyday colloquial phrases. The elicitation corpus was designed to represent linguistic structures.

Texts were translated by a combination of professional translators and crowd-sourced translators. Further details on the translation process are contained in the documentation accompanying this release.

Phrase-level data (elicitation and phrasebook text) is presented as tab-delimited (.tab) files. Document-level data (news text and translations) is presented as XML files. All text data is UTF-encoded.

Samples

Please view these samples:

Sponsorship

This material is based upon work supported by the Defense Advanced Research Projects Agency (DARPA) under Contract No. HR0011-15-C-0123. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of DARPA.

Updates

No updates at this time.

Available Media

View Fees





Login for the applicable fee