Globalex @LREC2020

The GLOBALEX 2020 Workshop @ LREC will follow up on the successful GLOBALEX workshops at LREC 2016 (https://globalex2016.globalex.link/) and LREC 2018 (https://globalex2018.globalex.link/). It is organized by Globalex – Global Alliance for Lexicography (https://globalex.link/), with support from:

• ELEXIS (EU’s H2020-funded project European Lexicography Infrastructure, https://elex.is/)
• TIAD (Translation Inference Across Dictionaries shared tasks and workshops, https://tiad2020.unizar.es/, https://tiad2019.unizar.es/, https://tiad2017.wordpress.com/)

This third iteration of GLOBALEX workshops at LREC will focus on linking data from lexicographic resources and will highlight aspects related to the automated linking of content among different dictionaries and other lexical sources, in the aim of enhancing linguistic data generation, enrichment and reinforcement.

Linking lexicographic data sets to each other and to other lexical resources, and in particular the interoperability of lexicography with Linked Data (LD) methodologies, have been gaining substantial attention in recent years, becoming a subject of various projects for research by and collaboration between academia and industry, including support of the public sector. Most notably, the W3C community group on Ontology-Lexica [1] was established following the release of the lemon model, which constituted the first de-facto standard for representing ontology-lexica, with the mission to “develop models for the representation of lexica (and machine readable dictionaries) relative to ontologies” [2]. The ensuing OntoLex-lemon model [3], [4] has served since 2016 as the leading option for conversion of lexicographic data into LD, and has recently been updated with the lexicog module [5] released on 17 September 2019 [6]. This trend has been complemented since 2015 by relevant literature (e.g. [7], [8], [9]), conference papers (e.g. [10], [11], [12], [13]) and EU-funded projects ([14] and [15], [16]).

Besides a section including general research papers, the workshop will include two shared task tracks – one on linking monolingual data and the other on linking bilingual and multilingual data, as follows:

1. Monolingual Word Sense Alignment – in conjunction with a shared task conducted by ELEXIS.
   Task 1 will be evaluated on novel dictionary linking data developed by the ELEXIS project [15], which will cover linking for the following languages: Danish, Dutch, English, Estonian, German, Hungarian, Irish, Italian, Serbian, Slovene and Russian.
   More details can be found here: https://competitions.codalab.org/competitions/22163
2. Linking Bilingual and Multilingual Lexicographic Resources – in conjunction with the 3rd TIAD shared task
   Task 2 will host the 3^rd edition of the Translation Inference Across Dictionaries (TIAD) shared task, of which previous editions were co-located at Language, Data and Knowledge conferences [17], [18]. The aim is to explore methods and techniques for automatically generating new bilingual (and multilingual) dictionaries from existing ones in the context of a coherent experiment framework that enables reliable validation of results and solid comparison of the processes used. In particular, the participating systems will be asked to generate new translations automatically among three languages, English, French, Portuguese, based on known translations contained in the Apertium RDF graph [19]. The inclusion of other language pairs will also be possible for this edition.