From the 3rd to the 7th of July 2026, I had the opportunity to travel to San Diego, California, to attend the 64th Annual Meeting of the Association for Computational Linguistic (ACL), the most prestigious conference in natural language processing. I had the honor to present two papers: “Protecting multimodal large language models against misleading visualizations” and “Is this chart lying to me? Automating the detection of misleading visualizations”, co-authored with researchers of TU Darmstadt and KU Leuven. Both papers address the threat of misleading visualizations, i.e., charts that use deceiving techniques such as inverted axes to misrepresent their underlying data. Our research focused on detecting such deceptive techniques and mitigating their negative impacts on automated chart understanding using multimodal large language models.

The conference featured more than 2,000 papers, presented over several oral and poster sessions. This was a great opportunity to discover new interesting research directions, to catch up with international collaborators, and to brainstorm about new research projects.

Some personal highlights include the SurgeLLM and EvalEval workshops, which featured interesting orals, keynotes, and panel sessions. The keynote by computer scientist Noah Smith at the Big Picture workshop on how to decide research directions as a junior researcher was also very insightful. Beyond academic discussions, it was also a great opportunity to explore San Diego and its surroundings. The Balboa Park, La Jolla Cove, and the sunset cliffs were all great places to discover.

Many thanks to emergenCITY and Athene Center for funding my research.

About the Author

Jonathan Tonglet is a PhD student at UKP Lab, Department of Computer Science, at TU Darmstadt. He is also affiliated with KU Leuven in Belgium. His research interests include multimodal automated fact-checking, automated chart understanding, and computational social science. He received his Master Degree in Artificial Intelligence with a focus on speech and language processing from KU Leuven in 2023 and works for the emergenCITY program area Information (INF).