Embed linear constraints in the covariance function of the Gaussian process
Combine the Bragg-edge method with a Gaussian process model
Gaussian processes for reconstructing the internal structure of an object from limited x-ray projections
Reconstruction of strain fields with varying lattice spacing
Fast and robust optimisation routine relaxing the quasi-Newton assumptions
Efficient and numerically robust approach to evaluate double line integrals of the squared exponential covariance function
A practical approach that allows for deep kernel learning in problems with intergal measurements.
I am a PHD student (started Feb 2017) at the Division of Systems and Control within the Department of Information Technology, Uppsala University. My supervisors are Thomas Schön, Niklas Wahlström and Adrian Wills. The main focus of my research is probabilistic modelling primarly focusing on the Gaussian process and its applications.
2018-06-20, Linearly constrained Gaussian processes, Swedish control meeting (Reglermöte), KTH Royal Institute of Technology, Stockholm. Slides
2019-01-22, Probabilistic approach for tomographic reconstruction, Deep learning and inverse problems 2019 (DLIP2019), KTH Royal Institute of Technology, Stockholm. Slides
2019-07-12, Reconstructing the strain field using a tailored Gaussian process, Applied inverse problems 2019 (AIP2019), Institut Fourier, Grenoble. Slides
To this date, I have been involved in teaching the following courses: