Skip to main content
Leo Wanner

Leo Wanner

Universitat Pompeu Fabra


Leo Wanner earned his Diploma in Computer Science from the University of Karlsruhe and his PhD in Computational Linguistics from the University of The Saarland, Germany. Prior to joining ICREA he held positions at the German National Centre for Computer Science (GMD), University of Waterloo, the University of Stuttgart and the Pompeu Fabra University, Barcelona. As visiting researcher, he was affiliated with U of Montreal, U of Sydney, U of Southern California's Institute for Information Sciences, U Paris 7, Columbia University, and U of Augsburg. Throughout his career, Leo has been involved as Principal Investigator in numerous of national and European research projects. He has published 8 books and more than 200 peer reviewed papers. He is Associate Editor of the Computational Intelligence and Frontiers in AI, Language and Computation journals and serves as regular reviewer for a number of high profile conferences and journals in the field.

Research interests

Leo Wanner is working in the field of computational linguistics, teaching the computer to understand spoken and written natural language material,  to supply people with information that might be useful to them and to interact with people. His research areas include human-computer interaction (in particular, the design and realization of conversational agents that reveal social and cultural competence), automatic written and spoken language generation, automatic summarization of written material, data-driven parsing, information extraction, and, more recently, abusive language analysis, author profiling and the information structure-prosody interface. He is furthermore interested in lexicology and lexicography, and there, in particular, in the automatic recognition, representation and use of lexical idiosyncrasies (so-called “collocations”) by both native speakers and learners of a language. An important characteristics of his research is that it is multilingual.

Selected publications

Wanner L. et al. 2021, ‘Towards a Versatile Intelligent Conversational Agent as Personal Assistant for Migrants‘. In: Dignum F. et al. (eds) Advances in Practical Applications of Agents, Multi-Agent Systems, and Social Good. The PAAMS Collection. Lecture Notes in Computer Science, vol 12946. Springer, Cham.

– Pérez-Mayos; A. Táboas García; S. Mille; and L.Wanner 2021, ‘Assessing the Syntactic Capabilities of Transformer-based Multilingual Language Models’. In Findings of ACL.

– Espinosa Anke, L.; J. Codina-Filbà; and L. Wanner.  2021 ‘Evaluating language models for the retrieval and categorization of lexical collocations’. In Proc. 16th Conf. of the European Chapter of the Association for Computational Linguistics.

– Pérez-Mayos; R. Carlini; M. Ballesteros; and L. Wanner. 2021 ‘On the evolution of syntactic information encoded by BERT’s contextualized representations’. In Proc. 16th Conf. of the European Chapter of the Association for Computational Linguistics

– Fortuna, P.; Soler-Company, J.; Wanner, L. 2021, ‘How well do hate speech, toxicity, abusive and offensive language classification models generalize across datasets?‘, Information Processing & Management, 58, 3, 102524.

– Shvets, A.; Fortuna, P.; Soler-Company, J.; Wanner, L. 2021, ‘Targets and Aspects in Social Media Hate Speech‘, Woah 2021: The 5th Workshop On Online Abuse And Harms, , 179 – 190.

– Fortuna, P.; Perez-Mayos, L.; AbuRa’ed, A.; Soler-Company, J.; Wanner, L. 2021, ‘Cartography of Natural Language Processing for Social Good: Definitions, Statistics and White Spots‘, Nlp4posimpact 2021: The 1st Workshop On Nlp For Positive Impact, , 19 – 26.

– Pérez-Mayos, L.; M. Ballestero; and L.Wanner. ‘How much pretraining data do language models need to learn syntax?’ In Proc. 2021 Conf. on Empirical Methods in Natural Language Processing.

Selected research activities

In 2021, Leo has also been PI of 6 ongoing large scale European and several national research projects.

ICREA Memoir 2021