Uncertainties in Large Language Models

About the Project

These projects are open to students worldwide, but have no funding attached. Therefore, the successful applicant will be expected to fund tuition fees at the relevant level (home or international) and any applicable additional research costs. Please consider this before applying.

Large Language Models (LLMs) are remarkably capable yet can be dangerously unreliable-producing confident but incorrect outputs (“hallucinations”) and failing to signal when they don’t know. This project tackles that risk by developing Bayesian deep learning methods for transformers and fast, scalable inference techniques that let LLMs represent and communicate both data (aleatoric) and model (epistemic) uncertainty. At a high level, we aim to make uncertainty in controllable language generation, reasoning, and decision-making.

Specifically, we will explore Bayesian transformers and lightweight posterior approximations (e.g. variational approaches, ensembles, etc), and develop new, efficient sampling methods for this task. We will build evaluation suites around calibration and reliability (ECE, NLL, Brier, etc) and demonstrate benefits on safety-critical tasks in healthcare, engineering, and finance. The outcome will be more trustworthy LLMs, with practical tooling and APIs for uncertainty quantification, that enable safer deployment in high-stakes settings.

Informal enquiries can be made by contacting Dr Y Huang (yongchao.huang@abdn.ac.uk)

Decisions will be based on academic merit. The successful applicant should have, or expect to obtain, a UK Honours Degree at 2.1 (or equivalent) in Computing Science. A strong background in machine learning, maths/probability theory, and programming (e.g. Python) is highly desirable. Candidates should have a keen interest in foundational AI research.

We encourage applications from all backgrounds and communities, and are committed to having a diverse, inclusive team.

Application Procedure:

Formal applications can be completed online: https://www.abdn.ac.uk/pgap/login.php.

You should apply for Degree of Doctor of Philosophy in Computing Science to ensure your application is passed to the correct team for processing.

Please clearly note the name of the lead supervisor and project titleon the application form. If you do not include these details, it may not be considered for the project.

Your application must include: A personal statement, an up-to-date copy of your academic CV, and clear copies of your educational certificates and transcripts.

Please note: you do not need to provide a research proposal with this application.

If you require any additional assistance in submitting your application or have any queries about the application process, please don't hesitate to contact us at researchadmissions@abdn.ac.uk

Funding Notes

This is a self-funding project open to students worldwide. Our typical start dates for this programme are February or October.

Fees for this programme can be found here Finance and Funding | Study Here | The University of Aberdeen