I completed my B.Sc. in Engineering Physics and M.Sc. in Engineering Mathematics at Chalmers University of Technology, and wrote my master’s thesis in the Healthy AI Lab. Since then I have moved to TU Delft, the Netherlands, as a Ph.D. candidate to work on causal inference and machine learning.
In my thesis, we studied the effect of having auxiliary privileged data on the sample efficiency of predictive models. This data is called privileged since it is only available during training and not at model deployment, yet it may still contain much useful information that we do not want to throw away. In particular, we concentrated on the problem of predicting future outcomes, where the privileged information consists of time-series data and there is only some baseline data available at prediction time – a setting that often comes up in clinical applications.