CALL FOR PARTICIPATION

We are excited to announce the CRAC 2022 Shared Task on Multilingual Coreference Resolution! The task will be held at COLING 2022 on October 16–17.

TASK DESCRIPTION

The focus of the task is learning models for prediction of coreference across a range of typologically different languages: Catalan, Czech, English, French, German, Hungarian, Lithuanian, Polish, Russian, and Spanish. Both identification of mentions and clustering of mentions that refer to the same entity are included. Only identity coreference is considered in this shared task, i.e., more complex coreference-related phenomena such as bridging anaphora are not considered. Submissions by participants will be ranked according to a scoring criterion based on standard coreference evaluation metrics.

BACKGROUND

Recently, inspired by the Universal Dependencies initiative (UD), the coreference community has started discussions on establishing a universal annotation scheme and using it to harmonize existing corpora. The discussions at the CRAC 2020 workshop led to proposing the Universal Anaphora initiative. One of the lines of effort related to Universal Anaphora resulted in CorefUD, which is a multilingual collection of coreference data resources harmonized under a common scheme. The public edition of CorefUD 1.0 contains 13 datasets for 10 languages, all of which are included in the CRAC 2022 shared task.

EXPECTED TIMELINE

Development Phase:

* March 3, 2022 – train+dev data release, start of the development phase
* March 17, 2022 – baseline system release, evaluation script release

Evaluation Phase:

* June 1, 2022 – start of evaluation phase, test evaluation via CodaLab
* June 8, 2022 (AoE) – end of evaluation phase

Presentation Phase:

* July 11, 2022 – submission of system description papers to CRAC 2022
* August 22, 2022 – notification of acceptance
* October 16-17, 2022 – the CRAC 2022 workshop at COLING 2022

Please visit the website for more details: https://ufal.mff.cuni.cz/corefud/crac22