A PhD student position is open in Prof. Dr. David Ginsbourger's uncertainty quantification and spatial statistics group at the Institute of Mathematical Statistics and Actuarial Science, University of Bern, Switzerland.
The recruited PhD student will be supervised by Prof. Dr. David Ginsbourger and affiliated with his group. The group is involved in collaborations with national and international partners from various research fields. These include geosciences, with a history of collaborations around stochastic modelling in hydrogeology, geophysical inversion, and atmospheric sciences (with links to the Oeschger Center of Climate Change Research). Recently, the group has also been increasingly collaborating with the Institute of Social and Preventive Medecine and the University Hospital around statistical machine learning for biomedical applications.
The recruited PhD student in statistics will work on the sub-project “Questioning similarities for expert-informed statistical learning” part of the collaborative project “Perception in Statistics and Econometrics” funded by the University of Bern. Distance- and similarity-based methods (nearest-neighbour classification, phylogenetic trees, multi-dimensional scaling, etc.) are at the heart of many predictive approaches in statistical data science and machine learning. Yet, distance/similarity functions are often chosen off the shelf, without necessarily questioning their adequacy to the considered task. For instance, the Euclidean and Gower distances typically appear as default choices when dealing with real-valued and mixed continous-categorical covariables, respectively. While metric learning has arose in recent years as an automatic method to tune distances in distance-based machine learning, a number of issues remain open when it comes to diagnosing the suitability of prescribed and tuned distance functions to tasks of interests. In this thesis, we will pioneer the field with novel diagnostics tools and algorithms extending approaches from spatial statistics and beyond (including variography, cross- validation, and more) to efficiently choose from and combine expert-informed distance functions towards improved predictivity and uncertainty quantification.
The ideal candidate will have recently earned or be about to finish their Master's degree in statistics or neighbouring subjects with a strong mathematical component, a genuine interest in statistical data science and applications thereof, a taste for both theoretical investigations and numerical experiments, and solid programming skills.
The salary will be at the level foreseen by the University of Bern. There might be a possibility to be involved in teaching and/or consulting duties. The funding is secured for up to 48 months with the starting date of July 1st 2023 or as can be arranged by mutual agreement.
Applications should contain: (1) a letter in which the applicants describe their research interests and motivations, (2) a complete CV, (3) copies of relevant diplomas, certificates and transcripts of records, (4) an electronic version of a research work (Master's thesis or other scientific publication), (5) contact information of 2 – 3 references.
Applications and inquiries should be sent to Prof. Dr. David Ginsbourger at [email protected]