PhD student position at the Division of Statistics and Machine Learning (STIMA) at Linköping University.
Your work assignments
We are looking for a PhD student working in the intersection of generative machine learning and computational statistical inference. Generative models based on diffusion processes have emerged as a prominent approach to machine learning with impressive performance in many application domains. A well-known use case is for image generation (these models are the main workhorse for tools such as DALL-E and Stable Diffusion) but the same technology has also shown great promise in applications as diverse as probabilistic weather forecasting, biochemistry, and materials discovery.
Conceptually, training a generative model is similar to solving a conventional statistical learning problem. Guided by this similarity, the research focus of the current position is to answer the questions:
- Can we leverage recent advances in generative AI for solving statistical learning problems?
- Can we leverage state-of-the-art statistical inference methods for improving generative modeling?
We will address these questions through novel methodological research resulting in new machine learning models and computational algorithms. We will also work on applied research to demonstrate the usefulness of the new methods, with particular emphasis on the application domains listed above. This is made possible by our active collaborations with applied researchers and domain experts within all of these fields.
The project will be carried out in a collaboration between STIMA (main supervisor: Fredrik Lindsten, senior associate professor in machine learning) and the Division of Systems and Control at Uppsala University (co-supervisor: Jens Sjölund, jens.sjolund@it.uu.se , assistant professor in AI). We will strive for a tight collaboration between the groups, including regular meetings and research visits. As a PhD student in the project, you are expected to actively engage in the teamwork and contribute to this collaboration.