Teaching

My teaching experience includes serving as a teaching assistant at Paderborn University for courses in Text Mining, Computational Argumentation, and Com­pu­ta­tion­al So­ci­o­lin­guist­ics Seminar. In this role, I was responsible for designing assignments that incorporated both coding and theoretical questions, leading assignment discussions during lab sessions, and correcting student submissions using various tools, such as Jupyter Notebook Grader. The following are the courses that I took part in leading.

Courses

Text Mining. The course introduces students to the core concepts, methods, and practical skills necessary to analyze and extract knowledge from natural language text. It covers foundational topics in linguistics, empirical research methodology, and machine learning, and teaches both rule-based and statistical approaches to text analysis, including grammars, lexicons, clustering, classification, and regression. Through Python-based tutorials and bi-weekly assignments, students learn how to implement text mining algorithms, work with real-world datasets, and design and evaluate scientific experiments.

Computational Argumentation. This course introduces students to both the theoretical foundations and practical techniques for analyzing and generating arguments in natural language. It begins with basic concepts from argumentation theory and natural language processing, then progresses to advanced tasks such as argument mining (identifying argumentative units in text), assessing the quality and structure of arguments, and even generating new arguments. The course also explores real-world applications, including argument retrieval for search engines and intelligent assistants. Through lectures and hands-on assignments, students apply these techniques, combining linguistic insight with computational methods to build systems that can understand and synthesize argumentative content.

Com­pu­ta­tion­al So­ci­o­lin­guist­ics Seminar. This seminar explores how language varies with social factors by combining insights from sociolinguistics with modern NLP techniques. It introduces theoretical frameworks for understanding language–society interactions, then applies them through hands-on work on topics such as media/social bias, argumentation, and language learning. The seminar gives students the chance to cover academic skills like scientific presentation and literature research, preparing them to engage in rigorous interdisciplinary research

When leading lab sessions, I designed assignments that included both written and programming components. I fostered collaboration by encouraging students to form groups of up to three. To enhance their engagement, students were required to present their assignment solutions during the lab session. In seminar courses, students undertake a research project by selecting a topic and are initially provided with two to three foundational papers. Their task is to conduct further research, using the initial materials as a springboard to develop a comprehensive research theme. This involves identifying relevant related work, proposing potential categorizations for different research lines, critically assessing the strengths and weaknesses of each contribution, and suggesting avenues for future research. The culmination of their work requires both a formal presentation of their findings and the submission of a survey paper. Throughout this process, I maintain a supervisory role, offering guidance while simultaneously encouraging independent exploration and discovery.

Mentoring and supervision

I have also supervised various master’s and bachelor’s theses at Paderborn and Columbia Universities, with several resulting publications at renowned NLP and IR conferences (see list below). In supervising theses, my method involves initially providing a rough sketch of the research idea and a few relevant papers to guide the student’s initial exploration. I strongly encourage independent research and critical thinking, where students develop and refine the research direction based on their own reading and insights. Through regular meetings, students present their literature reviews and explain how their findings contribute to the research. Recognizing that different students have varying needs, I strive to adopt a flexible supervision style, offering more or less guidance based on individual preferences and requirements.

A selection of students I supervised

  • Timon Gurcke (Paderborn University, Master’s Thesis): Thesis resulted in the paper “Assessing the Sufficiency of Arguments through Conclusion Generation”.
  • Nick Düsterhus (Paderborn University, Bachelor’s Thesis): Thesis resulted in the publication “Extractive Snippet Generation for Arguments”.
  • Jonas Rieskamp (Paderborn University, Master’s Thesis): Thesis resulted in the publication “Generating Contrastive Snippets for Argument Search”.
  • Arkajit Dahr (Paderborn University, Master’s Thesis): Thesis resulted in the paper “Argument Undermining: Counter-Argument Generation by Attacking Weak Premises”.
  • Narutatsu Ri (Columbia University, Project Work): Project resulted in the publication “Latent Space Interpretation for Stylistic Analysis and Explainable Authorship Attribution”.
  • Nikhil Reddy Varimalla (Columbia University, Project Work): Project resulted in the paper “Generalizable Analysis of Human Authorial Style by Leveraging All Transformer Layers”
  • Vishal Anand (Microsoft, Collaboration Project): Project resulted in the paper “iBERT: Interpretable Style Embeddings via Sense Decomposition”.