What's
Apr 6. 2026
KInIT team at EACL 2026
This spring, six of our colleagues travelled to Rabat, Morocco, to represent KInIT at the 2026 Conference of the European Chapter of the Association for Computational Linguistics (EACL 2026). We are proud to have had such a strong presence at one of the top NLP conferences, and even more so to let them tell the story themselves. Here are a few highlights from the conference, straight from our team.
Our papers presented at EACL 2026
Our research at EACL
For our PhD student Ivan Vykopal, the conference was a double experience. He not only presented two posters on multilingual NLP and fact-checking, but also joined as a volunteer and got to see how the conference works behind the scenes.
“Volunteering at the EACL 2026 Conference gave me the opportunity to meet many students from all over the world, see how a conference really works behind the scenes, and talk directly with attendees since I was there to help them. It was a wonderful experience, and I recommend every PhD student try it at least once during their studies.”
Presenting his work was just as important to him:
“Presenting my work at the conference was a real highlight. It was rewarding to see so much interest in my research and to feel that our work resonates with the research community. The questions and feedback helped me shape my ideas and reminded me why I enjoy doing research. Leaving the conference, I feel more confident that our research truly matters.”
Ivan Vykopal, PhD Student at KInIT
At the conference, our PhD student Róbert Belanec presented two papers, both dealing with the challenge of making LLM fine-tuning easier and more reproducible.
“During the conference, I presented two of our papers that aim to make large language model fine-tuning more accessible and reproducible. Our main paper, PEFT-Bench, introduces a unified benchmark for evaluating parameter-efficient fine-tuning methods across 27 NLP datasets and 7 PEFT methods, along with a new metric that factors in trainable parameters, inference speed, and memory usage. Our demo paper, PEFT-Factory, presents an open-source framework that supports 19 PEFT methods in a controlled environment for reproducible benchmarking, which has already drawn interest from researchers looking to incorporate it into an upcoming tutorial on adapters.”
Róbert Belanec, PhD Student at KInIT
Top moments from our researchers
One thing that came up more than once was the atmosphere of the conference itself.
“EACL 2026 was much smaller than other *CL conferences (such as ACL or EMNLP), but I saw this as an advantage. It allowed for deeper interaction with participants without time pressure, and made it easier to find interesting papers by simply walking through the posters. Overall, the conference felt more “intimate” and offered better networking opportunities, suggesting that smaller conferences may be more effective than large ones.”
Branislav Pecher, Researcher at KInIT
Meeting people in person was another big highlight.
“One of the highlights of EACL 2026 was the opportunity to network with new researchers and finally meet collaborators and project partners in person. The conference helped strengthen existing collaborations and foster new ones, potentially leading to future joint work. It was especially rewarding to connect with colleagues from DFKI, TU Berlin, and Charles University in Prague, and to see the poster on a browser-based verification plugin developed within the vera.ai project. Our own posters also sparked many discussions, alongside other presentations on improving fact-checking tools for media professionals.”
Martin Hyben, AI Specialist at KInIT
The venue and overall organisation left a strong impression, too.
“My experience at the EACL 2026 conference was very positive overall. The organisation was smooth and the venue was pleasant and easy to navigate, which made attending sessions stress-free and enjoyable. I especially appreciated the smaller poster sessions, which created a more relaxed atmosphere and gave me plenty of time to have meaningful discussions with other researchers rather than rushing from one poster to another. The social event was also very nice and helped foster a friendly and welcoming community feeling among attendees. Most importantly, I was impressed by the wide range of interesting research presented, particularly in the area of low-resource NLP, which featured many creative and impactful approaches.”
Ján Čegiň, Researcher at KInIT
“At EACL 2026, I co-authored six papers across key areas of natural language processing, including two on efficient training of large language models using PEFT techniques, developed with Robert, a PhD student I co-supervise. These works introduced new datasets, methods, experimental setups, and empirical insights that improve understanding of LLMs and support more reproducible research. Our group’s presence across multiple sessions made the research highly visible and led to many discussions within the community, as well as useful feedback for future work. Beyond the academic program, attending EACL in Rabat was also a valuable personal experience, offering a chance to explore the culture and atmosphere of Morocco’s capital.”
Ivan Srba, Researcher at KInIT
“EACL 2026 was a fantastic experience overall. The conference brought together researchers from institutes worldwide, with PhD students making up a large portion of the attendees. It was a great opportunity to exchange ideas, build connections, and get inspired about new research. A personal highlight was participating in a mentoring session with Yasmin Moslem from the ADAPT Center in Ireland, which ended with new ideas for my future research.”
Róbert Belanec, PhD Student at KInIT
And that is a wrap on EACL 2026! Conferences like this are where our research meets the world, and this one did not disappoint. We are proud of everyone who represented KInIT in Rabat, and we are already looking forward to the next one!
