In winter semester 2020 - 21, Jelena Mitrović is interim Assistant Professor of Computational Rhetoric and Natural Language Processing (NLP) at the University of Passau. She holds a master’s degree in information science with a focus on information retrieval and NLP and a master’s degree in Modern Greek with minors in ancient Greek and English. She conducted her PhD studies at the University of Belgrade, Serbia. In 2016, she started off as a junior researcher at the Faculty of Computer Science and Mathematics at the University of Passau. Since the early days of her PhD studies, Professor Mitrović has been interested in computational methods pertaining to the analysis of non-literal, figurative language, which also stems from her background in Greek and ancient Greek studies. To this end, she has developed an ontology of rhetorical figures that is the first of its kind, giving a formal description of 98 rhetorical figures from a syntactic and semantic perspective. Among other things, Professor Mitrović has developed a system for detecting ironic utterances in Twitter based on new semantic relations.
Is becoming a professor one of your lifetime goals?
Definitely! At the University of Passau, things have worked out quite well for me: While I have been appointed interim assistant professor, I have also successfully raised funding from the German Federal Ministry of Education and Research (BMBF) for a project I am going to work on for the next four years. The project is called “CAROLL – Computational Rhetoric in Social Media and Law” and it involves leading my own research group. My research combines the loves of my academic life: linguistics and computer science. In CAROLL, my team and I are going to incorporate deep learning techniques into a broad formal semantics approach, drawing insights from the ancient art of Rhetoric. We aim to develop a deep understanding of rhetorical arguments, implemented in various use cases, such as social media and the analysis of legal texts.
It all makes sense now, as computer science, linguistics and law are highly interconnected. They are all about logic, and we will investigate these connections in CAROLL.
Jelena Mitrović, interim Assistant Professor of Computational Rhetoric and Natural Language Processing
Algorithms that Aristotle would have liked: a team at the University of Passau is developing an Artificial Intelligence system based on the ancient art of rhetoric. This AI system could be able to detect implicit insults on the Web, but also to recognize which of two legal arguments is the better.
Are you the first professor in your family?
Yes, I am the first one. I was inspired by many strong women in my family, but also by my paternal grandfather. He was an impressive person who came from a challenging background, but managed to become a criminal judge. He was born in 1910. Being a criminal judge in the 20th century in a country with such an eventful history as Serbia definitely was a big deal. Because of him, my family was hoping that I would study law. However, law has never been interesting to me up until now. It all makes sense now, as computer science, linguistics and law are highly interconnected. They are all about logic, and we will investigate these connections in CAROLL.
Where are you originally from and what is your academic background?
Originally, I am from Užice, a small town in Serbia. Linguistics was my first love, so I studied Modern Greek Language in Belgrade. After I had my first child (who is now an 8th grade student at the nearby ASG secondary school), I decided to study computer science. My husband is a computer scientist, and he inspired this decision, and I’ve always liked mathematics and informatics, since primary school. In Serbia, I have the impression that girls in general are more encouraged to do mathematics and to go into STEM than it is the case in Germany. After finishing my computer science studies, I started to focus on Natural Language Processing, a subfield of computer science that has a strong grounding in linguistics and is one of the building blocks of AI. It seemed like a natural extension to what I had done before. I wrote my doctoral thesis on NLP for the Serbian language. In contrast to English or German, Serbian is a low resource language, which means that it is very difficult to train machine learning algorithms as there are not too many publically available tools and resources available. My work was aimed at overcoming these problems.
"I fell in love almost immediately with the picturesque town of Passau" - interim Assistant Professor Jelena Mitrović.
What made you decide to come to Passau?
Germany came as a natural answer as there are more job opportunities here and we can visit our families from here by car. Also, we had already had friends living in Passau. I heard of a job vacany at Professor Siegfried Handschuh's chair - he used to hold the Chair of Digital Libraries and Web Information Systems at the University of Passau, was looking for people. So, I applied, had a skype interview and got the job! When we arrived here in 2016, I fell in love almost immediately with the picturesque town of Passau. However, the arrival had been challenging as my daughter was only two years old and I started off working as a full-time researcher without knowing the German language. But for us, Bavaria and Passau have turned out to be the perfect choices, as Passau provides a very family-friendly environment.
It must be very difficult to conduct research in a foreign language?
It definitely helps that English is the main language used in computer science, and that we write research papers almost exclusively in English. In my new project CAROLL, however, we use German law texts for our analysis, for example the State Examination at the end of the Law studies. That is a challenge for a non-native speaker, but I am looking forward to learning and understanding these texts at a deeper level. However, upon arrival to Passau, not-so-perfect German was not a big problem for my family and myself. Although it is a small town in Bavaria, Passau is used to tourists from all over the world. There are many people, especially in the Old Town, who are able to speak English.
How do you perceive the atmosphere at the University of Passau?
I like the way people here at the university work: everything from the very beginning has been nice and collaborative. I also appreciate very much that interdisciplinary collaboration seems to be a natural order of things here at the University of Passau. It profits from its small campus. At other universities you don’t have that physical closeness.