- Paper submission deadline (extended): 13 February 2021
- Notification of acceptance: 22 February 2021
- Camera-ready due: 1 March 2021
- Workshop: 20 April 2021
Theme and Motivation
The languages from the Balto-Slavic group play an important role due to their diverse cultural heritage and widespread use — with over 400 million speakers worldwide. The current political and economic developments in Central and Eastern Europe bring Balto-Slavic societies and languages into focus in terms of rapid technological advancement and expanding consumer markets.
This year, we are especially glad to have an opportunity to organize BSNLP in a Slavic-speaking country.
The linguistic phenomena specific to the Balto-Slavic languages — such as rich morphological inflection and free word order — make the construction of NLP tools for these languages a challenging and intriguing task.
The goal of this Workshop is to bring together researchers from academia and industry working on NLP for Balto-Slavic languages. In particular, the Workshop aims to stimulate research and foster the creation of tools and resources for these languages. The Workshop will provide a forum for exchanging ideas, discussing current problems, and making the available resources more widely known. One fascinating aspect of this language group is the striking structural similarity, as well as an easily recognizable core vocabulary and inflectional inventory spanning the entire group of languages — despite a lack of mutual intelligibility — which creates a special environment in which researchers can appreciate the shared problems and solutions, and communicate naturally. This Workshop continues the proud tradition established by the 7 previous BSNLP Workshops.
This Workshop addresses Natural Language Processing (NLP) for the Balto-Slavic languages. The NLP tasks in urgent need of attention include:
- morphological analysis and generation,
- morphosyntactic tagging,
- syntactic and semantic parsing,
- lexical semantics,
- named-entity recognition,
- text normalisation and processing non-standard language
- coreference resolution,
- information extraction,
- question answering,
- information retrieval,
- text summarization,
- machine translation,
- development of linguistic resources,
- disinformation detection,
- fact verification.
Shared TaskThis year's BSNLP features the 3rd edition of the Shared Task on Multilingual Named Entity Recognition in Slavic languages, in recognizing mentions of named entities in Web documents in Slavic languages, lemmatization, and cross-language matching. The shared task covers Bulgarian, Czech, Polish, Russian, Slovene, and Ukrainian. Information about the Shared Task, as well as training data, is available on its web page.
We invite two types of submissions: long and short papers.
Long papers should describe original, unpublished and completed work. The short papers should describe: (a) work in progress and/or small focused contributions, or (b) system demonstrations, new linguistic resources, or experience with using existing software and resources, or (c) ongoing projects and activities that are relevant to stakeholders in the domain of Balto-Slavic NLP.
Overlap with previously published work should be clearly mentioned at the time of submission. The authors should indicate along with their submission whether the paper has been submitted elsewhere, e.g., to the main conference. In particular, in case the paper has been rejected by the main conference, it should be indicated in the submission.
All submissions will be judged on correctness, novelty, technical strength, clarity of presentation, usability, and significance/relevance to the Workshop. Submissions will be reviewed by at least three members of the Program Committee.
The review of long papers will be blind. Long papers must not include the authors' names and affiliations. Self-citations and other references that reveal the authors' identity must be avoided. The short papers may be anonymous, but that is not required.
In particular, submissions describing systems, resources, or solutions that have been made available to the wider public are strongly encouraged, as this will help promote computational linguistics applications for the Balto-Slavic languages.
Long papers should follow the two-column format of EACL 2021 proceedings not exceeding eight (8) pages of content plus two (2) pages for references. Short paper submissions should follow the same format, and should not exceed five (5) pages for content plus two (2) pages for references.
All submissions must conform to the official style guidelines of EACL 2021 contained in the style files and must be in PDF. Camera-ready versions of accepted papers must be provided both in LaTeX and PDF format.