Large volumes of high quality imaging data are being acquired by ongoing large astronomical surveys such as Sloan Digital Sky Survey or Dark Energy Survey. Even more, there are some upcoming research projects, like Large Synoptic Sky Survey, observing in optical wavebands, or eROSITA, the future X-ray satellite, which are expected to detect a lot of currently unknown objects, and also galaxy clusters among them; and galaxy clusters are a subject of matter of the proposed project.

The acquired data will provide an invaluable information for deep understanding of the Universe. Future projects should study galaxy clusters themselves, their evolution, and subsequently, giving constrains on some cosmological parameters. First step towards such tasks is to find and to classify galaxy clusters, moreover without any human touch. To accomplish these aims, solid skills in advanced machine learning techniques are required.

The student should familiarise with available multi-wavelength observations, sky surveys and methods commonly used in astronomy. The student will develop a method based on convolutional neural networks, to recognise galaxy clusters by their appearance on regular sky frames in distinct spectral bands. Some advanced techniques, like transfer learning, or citizen science project frameworks, are supposed to be utilised.

Thesis adviser will be Mgr. Maggie Lieu, PhD.