LT3 Newsletter 2025

We start by sharing the new projects that were launched in the past year:

LangTech	Rainbow - Researching Stereotypes towards LGBTQIA+ Community Members with Multilingual Natural Language Processing CLAIM - Towards Complex Reasoning in LLMs
TransTech	FWO bursary project on The influence of linguistic context and language similarity on structural priming: a multi-methods study

With the launch of these projects, we strengthened the LT3 team with a postdoctoral researcher (Alessandra) and two new PhD students (Francesca and Marzieh).  In 2024, we also received 2 HPC Tier 1 grants, allowing us to take our research projects to the next level with access to extra computing power. 

In Spring 2024, Suna joined our team as a visiting researcher for three months and in September we were also happy to welcome three new visiting/affiliated researchers: one postdoc (Liesbeth) and two PhD students (Delu and Shijie). During that same time Loic took up a new role as postdoctoral research assistant. Unfortunately in that same month we also had to say goodbye to Paola, whose Marie-Curie project DUAL-T came to an end. As a final event the workshop Literary Translators and Technology was organised which focused on user-centered research in literary translation and literary translators’ perspectives on technology. The workshop was a great success and we wish Paola all the best with her new position at the University of St. Andrews! 

Four junior researchers defended their PhD in 2024, a wonderful achievement!

Loic’s dissertation covered the challenges of coreference resolution in Dutch, with a special focus on event coreference resolution. Jasper defended his work on using NLP for vocabulary selection for learners of Spanish as a second language. Both defenses were preceded by a designated workshop allowing us to give the floor to experts in both fields.

• De Langhe, L. (2024). Demystifying discourse in Dutch : a study on event coreference resolution. Ghent University. Faculty of Arts and Philosophy (Ghent, Belgium).
• Degraeuwe, J. (2024). IVESS: Intelligent Vocabulary and Example Selection for Spanish Vocabulary Learning. Ghent University. Faculty of Arts & Philosophy (Ghent, Belgium).

Jonathan and Lilia, who were both affiliated with LT3, worked on an FWO infrastructure project to investigate the use of serious gaming to support syntactic annotation of a corpus of Spanish dialects. 

• Segundo Diaz, R. L. (2024). Juegos del español : iterative design, evaluation and implementation of games with a purpose to enhance parts-of-speech tagging in a corpus of European Spanish dialects. Ghent University. Faculty of Arts and Philosophy ; Hasselt University, Hasselt, Belgium ; Ghent, Belgium.
• Bonilla, J. E. (2024). Universal dependencies for spoken Spanish. Ghent University. Faculty of Arts and Philosophy, Ghent, Belgium.
  
  

Conference season was especially busy this year, with many team members being involved in the organisation of conferences, workshops and shared tasks. 

• Veronique was the programme chair for LREC-COLING, which took place in Turin. 
• Also at LREC-COLING, Els and Veronique co-organised the FinNLP-KDF-ECONLP workshop. 
• Sofie co-organized the tutorial Before Model Training: Crafting High-Quality Text Annotations for Machine Learning at ICWSM 2024.
• Our TransTech team was well-represented at EAMT2024 as each workshop at the conference was co-organised by at least one of our colleagues:
  • Joke & Janiça were involved in GITT 2024
  • Lieve & Paola co-organised CCT
  • Arda was involved in the first edition of KEMT
• Closer to home, Els & Colin organised the DAAL workshop, which was part of the Paratexts in Premodern Writing Cultures conference. 
• Veronique gave a keynote at JEP TALN Recital in Toulouse entitled “Towards a Reappraisal of the linguistic perspective in the LLM Era”.
• Orphée co-organised the WASSA2024 workshop at ACL 2024, which included the EXALT shared task organised by Els, Pranay & Aaron.
• Alessandra gave a keynote  at the Third Latin American Symposium on Research and Innovation Women in Engineering (III SIILMI) entitled “Language, Technology, and Society: Advancing Inclusion Through NLP”

Apart from the organisation, our researchers also presented their work at many different conferences and workshops throughout the year: LREC-COLING 2024, NITS, Anéla, SIG Writing, EAMT 2024, EuroSLA33, TaLC2024, ACL 2024, CLIN34, NLP4CALL, CLARIN 2024, EMNLP 2024, LW Research Day and 3° SIILMI.

Please refer to LT3’s 2024 publication list if you want to learn more about our latest research.

Besides conferences, we were also pleased to welcome various international guests at our LT3 offices in Ghent to discuss ideas and get inspired: 

• Steffen Eger and Ran Zhang from the NLLG lab came to present their work on the evaluation of text generation and digital humanities.
• Dan Zeman came over for a discussion with the group on his research domain of Universal Dependencies and syntax.
• Janiça also organised a fireside talk with Stefan Miedzianowski (Vice President Research at Deepl) about the state of the art and future directions of Machine Translation and Large Language Models. 
• Anastasia Giachanou gave a seminar called “Sexism, Fake News, and Beyond: Analyzing Online Behavior Through Language”. 

 

Looking back at 2024, there are some more achievements that fill us with pride:

• Veronique was appointed the prestigious Francqui chair at the ULB for the academic year 2023-2024. Moreover, in November she was also selected as a member of the Royal Flemish Academy of Belgium for Science and the Art.
• After a lot of hard work behind the scenes we officially launched our book Taaltechnologie Ontrafeld on June 18th! With this book, we aim to make language technology accessible and understandable for a broad audience. This book would not have been possible without all the hard work Cynthia put into it, together with her co-authors Veronique, Els, Lieve, Orphée, Arda, Joke and Seza. 

• The first LT3 spin-off AlfaSent was officially founded by Ellen, Veronique & Els! AlfaSent helps companies in becoming more customer-centric by using AI to analyse customer feedback. We were also featured in an article on DurfDenken.

 

The new academic year also meant the launch of our new language technology track in the Applied Linguistics bachelor! Instead of choosing 2 foreign languages, students can now choose 1 foreign language and a focus on Language Technology. After a communication campaign 15 students enrolled in this new track and we are hoping to keep that momentum going over the next few years. Please help us to spread the word!

This academic year, 13 students are following the CALM postgraduate programme. One of the main strengths of this programme is our collaboration with trusted partners, such as Yamagata Europe, where our students were invited for a company visit during the fall semester. 

We were also involved in another edition of the Methods in Language Sciences summer school. 

 

AI is present in daily life, education and work, and our outreach activities definitely confirmed this. Throughout the year various LT3 members were invited to share their insights on AI and LLMs for both education and companies.

Other outreach activities included:

• Joke hosted two webinars for KTV Kennisnet, called ‘Genderaanduidingen in het Nederlands & in vertaling' and together with Lieve they also organised another one on Machine Translation & Post-Editing.
• Veronique was part of a panel conversation about “taalkracht als werkkracht”.
• Cynthia gave a talk about subtitles as digital tools for language professionals. 
• Els gave a course at UGain about future proof data management. 
• Colin gave a guest lecture at King's College London on NLP for historic languages

 

Jasper created a number of useful resources for his PhD that are now publicly available

• WSD for ICALL: allows you to perform word sense disambiguation for Intelligent Comuter-Assisted Language Learning (ICALL) purposes.
• Example Sentence Selector: select pedagogically suitable example sentences from a set of sentences and rank them according to a degree of pedagogical suitability.
• LexComSpaL2: train personalised word-level difficulty classifiers for learners of Spanish as a foreign/second language.

 

All work and no play… let’s say that at LT3 we again did our best in 2024 to limit the dull moments 😉
This ranged from creative PhD Gifts:


   

To another round of Office Olympics for our summer teambuilding

To sightseeing and having fun during conferences