Although computational linguistics carries the promise of producing tools for processing and understanding a wide variety of languages, most of the work in NLP still focuses on a small number of languages, and in particular on English. The goal of this special issue is to promote linguistic diversity in NLP, by encouraging the publication of work on languages or language varieties less often studied, as well as methods that can easily and demonstrably be applied to those.
The challenges faced range from the low availability of resources for developing and estimating models, to difficulties in handling more complex morphology and syntax, to the steep publication curve for this type of work.
For this special issue, the "Revue TAL" journal encourages contributions (in English or French) describing, for example:
- new methods with a demonstrated application to less studied languages, dialects or language varieties;
- new applications of techniques and models, even already published ones, to new languages and language varieties, with a clear identification of the challenges;
- models of linguistic phenomena that are not, or less present in English and most commonly processed languages, as well as model or tools that implement them;
- producing or porting resources for less studied languages.
We also encourage the submission of contributions dealing with processing multilingual texts, for example for machine translation or learning multilingual or cross-lingual representations, when the languages under study increase the linguistic diversity.
Authors who wish to submit their work to this special issue are strongly encouraged to clearly identify the languages their work focuses on, as well as the specific challenges of the languages under study.