Abstract: 

In recent years the amount of astronomical data produced by telescopes has largely increased and it is expected to rise even further in the near future. This has resulted in machine learning techniques becoming more and more popular as a tool to analyze and study astronomical data. In particular, generative adversarial networks are unsupervised learning models that can learn a global distribution, such as that describing galaxy images. I will explain how GANs can be used to detect anomalies in astronomical data.