LT3 Newsletter 2024

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

Our team grew quite a bit in the last year. First, Florian joined our team, working on the DoD project. Later on, we welcomed 4 new PhD students: Natalia, Janiça, Joni and Thomas. On top of that, we also have a new professor (Prof. dr. Arda Tezcan) and a new affiliated professor (Prof. dr. Gilles-Maurice de Schryver.

In September, Olha successfully defended her PhD on sentiment analysis and emotion detection using fuzzy rough methods. Congratulations, once again, dr. Kaminska, we wish you all the best with your future endeavors.

• Kaminska, O.. (2023). Fuzzy rough set methods for emotion detection and sentiment analysis. Ghent University. Faculty of Sciences, Ghent, Belgium.

In 2023, we attended conferences all over the world to present our research. Here is an overview of all the conferences and workshops where LT3 researchers presented their latest work: EACL 2023, WASSA and CODI workshops at ACL 2023, CLIN33, RANLP, CLARIN, EAMT and the GITT workshop, EuroCALL, LSB Linguists’ Day, the CRAC workshop at EMNLP and EuroSLA.

On top of that, LT3 was involved in the organisation of a shared task on Multilingual Grammatical Error Detection at NLP4CALL 2023, the 13th WASSA workshop at ACL2023 and the GITT workshop at EAMT2023.

Moreover, the OPERAS project Machine Translation Evaluation in the Context of Scholarly Communication came to an end with this interesting publication

• Veronique is Programme Chair for LREC-COLING 2024.
• Janiça and Serafina were both granted FWO funding. Janiça will do research on automatic detection of (potential) factors in source texts leading to gender bias in machine translation. Serafina’s research will focus on automated writing evaluation for Dutch by employing LLMs.
• Ellen participated in Expedition DO! with her business idea for AlfaSent: an AI based software that helps companies keep track of their customers’ opinions using fine-grained sentiment analysis and emotion detection. She made it all the way to the finale.
• We have been able to welcome two international researchers at LT3 in 2023:
  • Postdoctoral researcher Alessandra Cignarella who worked on the homogenization of irony corpora in multiple languages (Italian, English and Dutch) together with Els thanks to an FWO International Mobility Scholarship.
  • Predoctoral researcher Johannes Sibeko who worked on readability assessment of the indigenous South-African language Sesotho together with Orphée thanks to funding from the UGent Global Minds Fund.
• At the same time LT3 members have also been conducting research abroad:
  • Pranay received funding for a research stay at the HiTZ Centre at the UPV EHU. Together with Eneko Agirre he worked on ideas for improving low resource language modeling through existing models.
  • Paola started doing experiments with literary translators at the Universiteit Leiden Centre for Linguistics. You can read more about it in their blogpost.
• In 2023, we also had to say goodbye to Bram, who is now a postdoc researcher at KU Leuven and Luna, who is now a Tenure Track Assistant Professor at UAntwerpen. We wish them both the very best and look forward to future collaborations.

Last summer, we were once more involved in the organisation of the MILS Summer School (Methods in Language Sciences).

In September we welcomed 10 new students for our postgraduate programme CALM. On top of that, 5 EM TTI students are finishing their second year with us. To create a true connection with the industry, we organised a company visit to Yamagata Europe. On top of that, we had three visiting scholars: Victor Sánchez, Eline Vanbiervliet and Lynne Bowker.

Finally, we have been working really hard on creating a new education track focused on language technology at our department. This track will launch next academic year.

At LT3 we do not solely focus on research, but also invest in informing the broader audience about recent developments in language technology and machine learning. To this purpose we launched the Dutch website ugent.be/taaltechnologie in 2023. In the same vein we have also been working very hard on writing the first Dutch accessible book on language technology, which will be published in Spring 2024.

We also gave several presentations and participated in a number of events:

• Veronique attended the Inspiration Day on AI at Hogeschool Gent.
• Els, Arda and Orphée gave lectures and presentations on Generative AI at Ghent University and for the Vlaamse Interuniversitaire Raad. Veronique gave a presentation at a study day for teachers.
• We organised an evening for teachers called ‘Language Technology Unraveled’, where Patrick, Arda, Orphée, Lieve and Cynthia presented the new developments in Generative AI, what machine translation entails and what the influence of language technology is on society.
• At Dag van de Wetenschap (Science Day), a number of our colleagues showcased our Furhat robot and other demos. On top of that Els and Paola both gave a lecture, Els participated in a panel on generative AI and Veronique presented the Burgerpraat project during a salon conversation.
• Veronique and Sofie spoke at the Ghent University alumni day.
• Patrick hosted a digital helpdesk for schools to answer questions about language technology.
• Els and Cynthia presented at the workshop ‘Spreken we morgen AI in de klas?’

 

The resources library on our website also grew in the past year, with an Explainer video on Neural Machine Translation and a quiz on Machine Translation. On top of that, the MATEO project came to an end, but we created a website where you can evaluate the quality of machine translation systems.

Finally, the DeBiasByUs website is creating a user-fueled database of bias in machine translation.

Finally, we’d love to share some pictures of the past year. Besides all the hard work, we also made time for some fun team building activities and multiple karaoke sessions.