Daily Image
05-02-2024
Astron/Leiden student Rafaël Mostert successfully defended his thesis
| Submitter: | Raffaella Morganti |
| Description: | On 25 January 2024, Rafaël Mostert successfully defended his thesis at the University of Leiden entitled 'Machine Learning for Radio Galaxy Morphology Analysis'. Rafaël's PhD was a shared ASTRON & Leiden University project. For his PhD project, Rafaël was supervised by Huub Röttgering, Raffaella Morganti and Ken Duncan. Rafaël's thesis has made a number of steps forward in the exploration of the use of machine learning to help astronomers in coping with the huge amount of images and sources to classify produced by the new radio telescopes. This is a fast expanding field and Rafaël has covered a number of relevant problems. One of the project part of the thesis was aimed at developing automatic ways to associate components from the source-finders into large radio sources. This is a necessary step in order to allow the optical identification and the study of the physical properties of the sources. His work has further expanded into the automatic classification of radio sources using self-organising maps (SOMs) in combination with a random forest classifier to reduce visual inspection when finding some of the most elusive radio AGN: remnants radio sources. The thesis covers also other topics and you can read all of them by downloading the pdf at: https://home.strw.leidenuniv.nl/~mostert/thesis/ Interestingly, the cover (designed by Rafaël) of the thesis (figure on the left) is an image created by a stable diffusion neural network hosted by stability.ai based on the prompt ‘A field of LOFAR antennas underneath a sky of radio galaxies, painting by Vincent van Gogh’. In reality, LOFAR antennas are not steerable dishes as depicted. Future generative AI models might correctly paint LOFAR antennas as static. Congratulations Rafaël! and good luck with your new job which may also bring many challenges! |
| Copyright: | ASTRON |
| Tweet |  |