Would you like to help us create large language models that can express their confidence in a way people understand?
We are offering an opportunity to pursue a fully funded PhD in computer science, specialising in data science/natural language processing, in the context of an innovative research project aiming to improve the trustworthiness of large language models (LLMs) and enable them to signal their confidence effectively to their users. You will be part of an exciting project spanning natural language processing, machine learning and experimental linguistics. You will develop methods to express uncertainty with large language models in a controllable way and align the level of confidence that is communicated with the user’s perceived confidence.
The project has a dual focus on advancing computational methods in natural language processing and validating them in user studies with human participants. The ideal candidate will have a strong interest in both parts of the project and a solid background in at least one of them.
The project is jointly led by Dr. Christian Hardmeier at ITU and Dr. Jes Frellsen at the Technical University of Denmark in close collaboration with Prof. Hannah Rohde at the University of Edinburgh. An extended research stay in Edinburgh is an essential part of the project.
Apply here: https://candidate.hr-manager.net/ApplicationInit.aspx?cid=119&ProjectId=... Deadline: 24 June 2026
Very soon, we'll also advertise a second PhD position in the same project at DTU, which will be more focused on machine learning. The advertisement will follow on this list!
-- Christian Hardmeier Associate Professor, IT University of Copenhagen https://christianhardmeier.rax.ch/