
Thomas Moerman PhD Researcher in NLP and AI, Ghent University (LT3)
I develop retrieval-augmented and synthetic-data methods for natural language processing (NLP) and artificial intelligence (AI), applied to a range of tasks such as machine translation, educational NLP, and fairness in language models.
News
- October 2026 Research stay at Transducens, Universitat d’Alacant, opened with a kickoff talk, Fuzzy Matches and Synthetic Data for Machine Translation.
- October 2026 Presented ContraGAND, a contrastive corpus for gender ambiguity detection, as a poster at EMNLP 2026 in Budapest. The paper is in the ACL Anthology, with the dataset, models and a demo on Hugging Face.
- October 2026 Our WMT 2026 submission for multitask Sorbian NLP won the Sorbian track at the Conference on Machine Translation (WMT), where we presented it as a poster in Budapest.
- September 2026 At CLIN 36 in Brussels, gave an oral presentation, Synthetic Data for Machine Translation, and presented the SanreMello Corpus poster, seven decades of Sanremo and Melodifestivalen as data.
- August 2026 Our EAMT 2026 paper on machine translation in the asylum context is now published in the ACL Anthology.
- July 2026 Taught the five-day Introduction to Python module at the MILS summer school (Methods in Language Sciences), Ghent University.
- June 2026 Presented our EAMT 2026 paper on machine translation in the asylum context, shortlisted for a best-paper award.
- May 2026 ShAnEL-2, our multilingual short-answer grading dataset, was accepted at LREC 2026.
Awards
- 2026 Winner of the Sorbian track of the WMT 2026 shared task on multitask LLMs with limited resources, as team lead of LT3. One 2B model handles translation, question answering, spelling, grammar and maths for Upper and Lower Sorbian.
- 2026 Best-paper award shortlist at EAMT 2026 for Multilingual Communication in the Asylum Context.