*MER-TRANS-2026: https://lastus-taln-upf.github.io/mertrans-iberlef-2026/
<https://lastus-taln-upf.github.io/mertrans-iberlef-2026/>IBERLEF 2026:
https://sites.google.com/view/iberlef-2026
<https://sites.google.com/view/iberlef-2026>Apologies for cross-postingWe
are pleased to announce the launching of the Shared Task: Multilingual Easy
to Read Translation (MER-TRANS) in the context of the IberLef 2026
Evaluation Forum. - Context*
*Linguistic access to information is increasingly recognised as a
fundamental citizens’ right. For example, the United Nations Convention on
the Rights of Persons with Disabilities (UNCRPD) includes accessibility as
one key enabler for a more inclusive society, while the European Union
adopted the European Accessibility Act (EAA) in 2019, which requires that
many everyday products and services, including digital information
services, comply with accessibility standards. The UNCRPD and the EAA
stress that information has to be provided in accessible language, such as
plain language or easy-to-read language in order to allow the different
population segments with language comprehension difficulties to exercise
their recognized right to participate in society and public life. Providing
information in easy language in a real-world setting is challenging.
High-quality, easy-to-read content must typically be written or translated
by human experts following specialized guidelines, and then often validated
by target readers. This manual process requires considerable effort, time,
and expertise, which limits how much content can be produced or translated
into easy formats. Therefore, this MER-TRANS shared task aims at advancing
the state of the art in the field of automatic easy-to-read translation
with specific emphasis on Romance languages, more concretely, in Catalan,
Italian, and Spanish. *
* - Task OverviewIn this shared task, we invite participating teams to
automatically produce easy to read versions of texts and/or sentences. The
texts will be sampled from the iDEM corpus, a multilingual corpus in the
domain of democratic participation, which has been simplified by human
experts following easy-to-read recommendations and high-quality validation
procedures. There will also be a surprise language task to be revealed
closer to the release of the test data. Up to three submissions per
language will be allowed per participant team. - Data and EvaluationUnlike
previous challenges where the data remained of restricted use, the iDEM
corpus, with original and adapted versions, will be made fully available to
the community, the documents for participants to produce adaptations will
be provided during the test phase, and the full dataset during the IberLef
workshop in September 2026. A small trial dataset will be released; it will
be sampled from the iDEM corpus, considering the occurrence of different
difficulties addressed by the easy-to-read adaptations. The examples will
contain both the original text excerpts and easy-to-read versions. The test
datawill consist of only the original complex text excerpts. Evaluation
will be carried out with automatic metrics borrowed from current text
simplification evaluation research, such as SARI and/or BLEUE and semantic
similarity scores when appropriate. - System Description Papers and
ProceedingsAll participating teams will be invited to submit a system
description paper describing their methods, models, and experimental
findings: further information about formatting and length will be given in
due time. Submitted papers will be reviewed by at least two peer
reviewers.Papers will be required to describe fully Reproducible solutions
which contribute to Open Science. Accepted system description papers will
be published at no cost in the Proceedings of the Iberian Languages
Evaluation Forum (IberLEF 2026), hosted by CEUR Workshop Proceedings
(CEUR-WS.org).Authors of accepted papers will be invited to present their
work at the IberLEF 2026 workshop. Presentation at the workshop is
encouraged but not mandatory for publication in the proceedings. - Who
Should Participate?Participation is open to all. - Important
DatesRegistration opens: Open nowRegistration closes: 28 Feb 2026Trial
data release: 06 Mar 2026Test data release: 06 Apr 2026System
outputs submission deadline : 13 Apr 2026Evaluation results published:
27 Apr 2026System Description Papers: 01 Jun 2026Papers acceptance:
14 Jun 2026Camera-ready papers due: 21 Jun 2026IberLEF 2026 Workshop:
22 Sept 2026 - How to ParticipateIn order to participate, teams must
register using the registration form available at the task website
https://lastus-taln-upf.github.io/mertrans-iberlef-2026/
<https://lastus-taln-upf.github.io/mertrans-iberlef-2026/>Note that
registration is mandatory in order to access the data and submit system
outputs. - Task OrganizersHoracio Saggion — Universitat Pompeu Fabra,
SpainNelson Pérez Rojas — Universidad de Costa Rica, Costa RicaStefan Bott
— Universitat Pompeu Fabra, Spain Nouran Khallaf — University of Leeds,
UKMehrzad Tareh — Universitat Pompeu Fabra, SpainDaniel Adanza —
Universitat Pompeu Fabra, SpainAlmudena Rascon — Plena Inclusion Madrid,
SpainSandra Szasz — Universitat Pompeu Fabra, Spain*
--
Horacio Saggion
Full Professor / Chair in Computer Science and Artificial Intelligence
Head of the Natural Language Processing Group - TALN
Project Coordinator iDEM Project (HE)
Co-PI of the AI-BOOST project (HE)
PI of the IDEAL project (HE)
Universitat Pompeu Fabra
https://twitter.com/h_saggion
https://www.linkedin.com/in/horacio-saggion-1749b916
--
Horacio Saggion
Full Professor / Chair in Computer Science and Artificial Intelligence
Head of the Natural Language Processing Group - TALN
Project Coordinator iDEM Project (HE)
Co-PI of the AI-BOOST project (HE)
PI of the IDEAL project (HE)
Universitat Pompeu Fabra
https://twitter.com/h_saggion
https://www.linkedin.com/in/horacio-saggion-1749b916
_______________________________________________
Corpora mailing list -- [email protected]
https://list.elra.info/mailman3/postorius/lists/corpora.list.elra.info/
To unsubscribe send an email to [email protected]
