I always enjoy supervising motivated students during their theses or internships. Below are a few example topics and areas that I’m currently very interested in or that would be well-suited for a thesis. I am always happy to mentor motivated students on their theses or internships. Below, I have listed a few sample topics and areas that are currently of great interest to me or that would be well-suited for a thesis. In addition, I provide an overview of the process - from initial contact through supervision to the final submission of the thesis. Individual details or questions can be discussed during an initial meeting.

Potential thesis research fields:

  • Annotator Variation - Investigating differences between human annotators and understanding when such variation reflects noise, ambiguity, subjectivity, or meaningful differences in perspective. Possible directions include inter- and intra-annotator agreement, disagreement analysis, annotation guidelines, consensus formation, perspectivist approaches, and modeling individual annotator behavior. It could also cover investigations whether individual annotators have systematic tendencies or biases and whether these can be modeled or predicted from their previous annotation decisions. Furthermore, implications for evaluation and model development could also be covered. References (ordered chronologically): 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, …
  • Narrative Extraction - Investigating methods for automatically identifying, extracting, and representing narrative structures from text. Possible directions include the extraction of events, entities, temporal and causal relations, characters, actions, and other narrative elements, as well as modeling how these elements are connected to form coherent stories. This could cover rule-based, supervised, and LLM-based approaches, comparisons between different extraction strategies, and investigations into how narrative representations can support downstream tasks. Furthermore, a potential thesis could explore different levels of granularity of narratives or how narrative structures vary across genres, domains, languages. In addition, it could could cover how extracted narratives can be evaluated (which would potentially touch the Annotator Variation field).
  • Event Detection - Quite similar to Narrative Extraction. Possible directions include detecting event mentions, identifying event types and arguments, distinguishing relevant events from background information, and modeling temporal, causal, or other relations between events.
  • …

The process described below applies to the supervision of master’s theses. For bachelor’s theses, all points can be adapted accordingly, i.e. presentation approximately 3 weeks before the submission deadline; presentation no longer than 20 minutes + 15 minutes of discussion. All points also apply mutatis mutandis to the supervision of internships, i.e. presentation approximately 1–2 weeks before the end of the internship; presentation no longer than 20 minutes + 15 minutes of discussion.

Upon initial contact/inquiry, please send me the following information:

  • transcript of records
  • either: an exposé (less than 1 page) explaining the proposed project (e.g., please submit this for external master’s theses) or, alternatively, at least a brief (typically a few sentences to less than one page) statement of motivation/explanation of your interests (the scope of potential supervision is not limited solely to the projects described on my website)
  • Desired start date/latest start date (e.g., “as soon as possible”)
  • Desired submission date/latest submission date (e.g., “by the submission deadline”)
  • Preferred primary supervisor/secondary reviewer
  • If relevant: any known restrictions on when the thesis must be corrected, if applicable

Procedure:

  • The student submits a request to the examination office and receives a registration form as confirmation that all requirements have been met (please refrain from contacting me until you have received this form).
  • Initial contact following the template above
  • Topic selection/definition (should typically take only a few weeks); this can be done (depending on needs and topic) entirely via email or through one or two brief meetings (max. 30 min each).
  • The student prepares an overview of the objectives and planned methods according to ADEMP (see paper, Section 3 with all subsections). All ADEMP elements must be specified and submitted to me once before registration.
  • Register the topic with the examination office using the registration form (the student sends the PDF to the examination office)
  • I will not force anyone to meet with me. Once registered, the student is responsible for successfully completing and submitting the thesis. This also includes asking me for meetings as needed (please indicate preferred dates directly and specify whether you prefer to meet digitally or in person at my office).
  • After approximately 5 months (or approximately 1 month before the submission deadline), a presentation of all results achieved up to that point is mandatory. The presentation itself should not exceed 30 minutes. The meeting is scheduled for 45 minutes. Please set aside 60 minutes in case more discussion is needed.
  • Additionally, I recommend sending me a draft of at least the planned table of contents once (no later than one month before the submission deadline).
  • I strongly recommend against submitting your thesis on the last day of the deadline. In general, it is helpful for the quality of your thesis if it is completed at least 7 days before the deadline. Read through your entire thesis after letting it sit for 2 days without editing, and correct only (!!) spelling errors. You should avoid making any significant changes to the content this close to the submission deadline. Upload your thesis in a timely manner (according to this calculation, a few days before the deadline).
  • Once submitted, the thesis will be graded within 3 months.
  • I usually send you a version of the thesis with comments immediately after grading it. This does not include a systematic scoring rubric and is not legally binding in terms of the examination, but is intended as an additional service to help you better understand the evaluation and identify any remaining weaknesses in your work that you should address for your future professional or academic career. If you have any questions about the comments, I am always willing to answer them in a constructive and substantive manner. However, I will not engage in discussions about individual comments with the aim of improving your grade.