Bachelor Theses
Important information
Here you find an overview about our currently available bachelor theses. If you have not found a topic that interests you, but are interested in writing your thesis at IMRG, please direct contact Prof. Dr. Schwabe
Please, keep in mind that thesis submitted in August and February will not be corrected before semester degree conferral dates!
Further information about the BSc program can be found at Bachelor in Informatics
Understanding Generative AI's Role in Academic Review Processes
Reserved. In this thesis project, the student will examine the role of generative AI on scientific reviewing in the information systems discipline. The research aims to understand how much and how AI-generated content affect scholarly peer reviews.
For more information, contact Dr. Liudmila Zavolokina
Integrating humor cues in AI propaganda detection tool to enhance critical thinking
Reserved. This thesis explores how integrating humor can enhance critical thinking in AI systems for detecting propaganda. It aims to identify effective ways to use humor to better understand and distinguish propagandistic content, fostering critical thinking among news readers. The student will design and evaluate the effectiveness of humor cues in an experimental setting to help readers better distinguish propaganda. The thesis is embedded in the on-going research project, see:Project description,Related reading
Interested? Contact Dr. Liudmila Zavolokina
Was denken Sie – warum haben Sie die Schmerzen? Was bedeuten die Schmerzen für Sie? Leveling up qualitative research using AI
In Zusammenarbeit mit der Abteilung für Chiropraktische Medizin an der Universitätsklinik Balgrist soll ein Chatbot entwickelt werden, der Patienten und Patientinnen mit chronischen Schmerzen des Bewegungsapparates nach ihrem persönlichen Verständnis ihrer Schmerzen, ihrer Selbstwirksamkeit sowie mentaler Bilder, die sie mit ihren Schmerzen assoziieren, befragt. Forschungsergebnisse deuten an, dass diese Faktoren sehr wichtig für die outcomes, die Lebensqualität der Patienten und letztendlich für den Behandlungserfolg sind. Allerdings leidet die qualitative Forschung bisher an kleinen Datensätzen, die durch menschliche Interviewer erhoben werden. Diese Lücke soll durch diese Masterarbeit geschlossen werden. Patienten zur Datenerhebung stehen bei diesem Projekt in grossem Umfang zur Verfügung.
Voraussetzungen: Gute Programmierkenntnisse, idealerweise mit Generative AI, gute deutsche Sprachkenntnisse
Startdatum: sobald wie möglich
Betreuung Universitätsklinik Balgrist: Prof. Petra Schweinhardt, Abteilung für Chiropraktische Medizin
Betreuung UZH Prof. Dr. Gerhard Schwabe
AI-controlled agents in virtual reality: prototype development of an immersive conversation simulator
Conversational skills are a key competence in many professional fields. Virtual reality (VR) makes it possible to train this interactively and situationally by providing an artificial agent with which a natural conversation situation can be practiced. In this thesis, a prototypical conversation simulation for specific communication situations will be designed, implemented and evaluated.
The focus of this thesis is on the development of a conversation engine. This should use the capabilities of existing speech-to-text, text-to-text, text-to-speech and possibly speech-to-speech models to enable natural conversation. Specifically, the development work includes connecting to the APIs of the selected foundation models and prompting the models accordingly. The costs for the interaction with the selected models will be covered. Furthermore, an interface to a VR engine will be specified and integrated. In addition, the thesis embraces the design of a flexible architecture and guidance on how to adapt to different scenarios. This may include, for example, the identification of prompting patterns that are effective in simulating consistent personalities. Finally, the prototype should be formatively evaluated and ideally published as an open source project.
The VR environment in which the created agent will be active and the virtual representation of the agent will be developed in a separate final project. A close collaboration with the student in this separate thesis is expected.
We are looking for students who like to explore and are willing to explore the capabilities of existing Foundation models. Good programming skills are very welcome. Willingness to collaborate with external partners is required.
For more information contact Mateusz Dolata