Antonio D'Isanto ( Italy )
Antonio.Disanto @ h-its.org
On the application of modified deep learning techniques in the era of terabyte-sized astronomical surveys
The data explosion of the recent years, due to the availability of the new synoptic surveys, has led to an increasing request of new methods of analysis in Astronomy. In particular, deep learning techniques are proving to be particularly efficient in the analysis of huge data-sets. This is valid for many different tasks, from classification to regression problems. Important examples of applications can be found in the determination of photometric redshifts or the study of time series and transient phenomena. Furthermore, the extension of this kind of techniques with the possibility of accurate error estimation constitutes a great step forward in many aspects of research in Astronomy. The main effort is focused on the realization of fully-automatized workflows that could process and analyze complex data to extract useful information and knowledge.
Supervisor: Kai Polsterer ( HITS )