Introduction: LORELEI Swahili Representative Language Pack consists of Swahili monolingual text, Swahili-English parallel text, annotations, supplemental resources and related software tools developed by the Linguistic Data Consortium for the DARPA LORELEI program. 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: Swahili is a Bantu language spoken by 100 million people throughout East and Central Africa. Significant populations of Swahili speakers can be found in Tanzania, Kenya, Uganda, and the eastern Democratic Republic of the Congo. Data was collected in the following genres: discussion forum, news, reference, social network, and weblogs. Both monolingual text collection and parallel text creation involved a combination of manual and automatic methods. Data volumes are as follows: Over 4.3 million words of Swahili monolingual text, approximately 409,000 of which were translated into English 90,000 Swahili words translated from English data 545,000 words of found Swahili-English parallel text Approximately 100,000 words were annotated for simple named entities, and up to 26,000 words were annotated for full entity (including nominals and pronouns), entity linking, simple semantic annotation, and situation frame annotation. Lexical resources and software tools are also included in this release. The tools recreate original source data from the processed XML material, condition text that data users download from Twitter, apply sentence segmentation to raw text, and support named entity tagging. Monolingual and parallel text are presented in XML with associated dtds. Annotation data is presented as tab delimited files or XML. All text is UTF-8 encoded. The knowledge base for entity linking annotation for this corpus and all LORELEI Representative Language and Incident Language Packs is available separately as LORELEI Entity Detection and Linking Knowledge Base (LDC2020T10). 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.

Creator(s)

Jennifer Tracey

Stephanie Strassel

David Graff

Jonathan Wright

Song Chen

Neville Ryant

Seth Kulick

Kira Griffitt

Dana Delgado

Michael Arrigo

Distributor(s)

Linguistic Data Consortium

Right Holder(s)

Portions © 2002-2007, 2009-2010 Agence France Presse, © 2000 American Broadcasting Company, © 2011-2017 BBC, © 2000 Cable News Network LP, LLLP, © 2008 Central News Agency (Taiwan), © 2007-2008, 2014-2017 China Radio International.CRI, © 2009, 2015 Dicasterium pro Communicatione, © 1989 Dow Jones & Company, Inc., © 2015-2016 Global Publishers, © 2017 Independent Television Limited, © 2017 Ipsinternational, © 2005 Los Angeles Times - Washington Post News Service, Inc., © 2015 Makuruki.rw, © 2014-2017 Millard Ayo, © 2014-2016 Mtanzania Newspaper, © 2014-2016 Nation Media Group, © 2000 National Broadcasting Company, Inc., © 1999, 2005, 2006, 2010 New York Times, © 2016 Pars Today, © 2000 Public Radio International, © 2015 Radio Rahma, © 2011-2016 RFI, © 2016 Tanzania Standard Newspapers, Ltd., © 2003, 2005-2008, 2010 The Associated Press, © 2015, 2017 TRT, © 2003, 2005-2008 Xinhua News Agency, © 2017 ZanziNews, © 2017, 2022, 2023 Trustees of the University of Pennsylvania