Special focus on multilingual NLG for under-resourced languages
We are delighted to announce a new edition of the WebNLG challenge, which will take place in 2022. WebNLG 2022 will focus on multilingual generation for under-resourced languages.
WebNLG
The WebNLG challenge consists in mapping data, in the form of RDF triples, to natural language text. The input is a set of RDF triples sourced from DBPedia for example:
(John_E_Blaha birthDate 1942_08_26)
(John_E_Blaha birthPlace San_Antonio)
(John_E_Blaha occupation Fighter_pilot)
where the corresponding output text might be:
John E Blaha, born in San Antonio on 1942-08-26, worked as a fighter pilot
The WebNLG challenge was launched in 2017. A second edition, in 2020, extended the task to Russian, in addition to English.
WebNLG 2022
This new edition of WebNLG focuses on four under-resourced languages which are severely under-represented in research on text generation, namely Maltese, Irish, Breton and Welsh. In addition, WebNLG 2022 will once again include Russian, which was first featured in WebNLG 2020.
Development and test data will be provided for all 5 languages. Data for each language was obtained by sourcing high-quality translations of the original English texts in the WebNLG 2020 dev and test sets. Training data will be provided for English and, as per WebNLG 2020, for Russian.
Timeline
The expected timeline for the task is as follows:
April 2022: Official announcement of the shared task and distribution of development data in all languages
September 2022: Evaluation of submitted systems on test data.
Results from the WebNLG 2022 challenge will be presented and discussed at a specially organised event in the final quarter of 2022.
Organisation:
WebNLG 2022 is being organised under the auspices of LT-Bridge, supported by the Horizon 2020 Work Programme Spreading Excellence and Widening Participation (WIDESPREAD) 2018-2020.
Claire Gardent, CNRS/LORIA, Nancy, France
Albert Gatt, Utrecht University, The Netherlands and University of Malta
Claudia Borg, University of Malta
Enrico Aquilina, University of Malta
Anya Belz, Dublin City University, Ireland
John Judge, Dublin City University, Ireland
Liam Cripwell, CNRS/LORIA and Université de Lorraine, Nancy, France
William Soto-Martinez, CNRS/LORIA and Université de Lorraine, Nancy, France
