We invite you to apply for a PhD position in applied mathematics at Lund University. Lund is a wonderful city to live in, its large student population will make you feel at home immediately.
The topic of the thesis will focus on the development of machine learning methods for anomaly detection and spatio-temporal analysis of mixed data including geospatial visible and multi-spectral signals. The research group will work with:

ways to describe correlations and noise in the data;
design of improved loss functions which enhance spatial-temporal analysis of information from remote sensing (GIS/satellites/drones);
design of improved autoencoders (VAEs, GANs), diffusion models and other such machine learning methods towards analysis and detection of anomalies as well as identifying important features of a learned probability distribution in latent space.

Analysis through information theory tools such as the relative entropy rate and the Fisher information matrix can be useful metrics in some of the above tasks. This work has diverse applications in image processing and analysis ranging from predicting yield to uncovering detailed carbon sequestration profiles over larger geographical regions.

More information: https://lu.varbi.com/en/what:job/jobID:626657/iframeEmbedded:0/where:4
or contact directly: carina.geldhauser@math.lth.se