The CTA CIEMAT group participates in the development of algorithms for real-time analysis of CTA data in collaboration with the Scientific Computing Unit of the Department of Basic Research. This part of the analysis chain is a key element to respond to opportunistic targets detected by other instruments in other energy bands and as a tool for making decisions during CTA observations.

Furthermore, the group also investigates the performance of advanced Machine Learning algorithms used for the reconstruction of CTA data. To understand the reconstruction of the data is necessary to know what a Cherenkov telescope image is. Cosmic rays interacting with the atmosphere generate electromagnetic shower along their paths, which produces short pulses of light collected by Cherenkov telescopes. The resulting images contain information about the properties of the particle that generated the shower. For Cherenkov telescopes science it is essential to distinguish if the electromagnetic shower was generated by a primary cosmic gamma rays or by the much more abundant primary hadronic cosmic ray. This is achieved by means of distinctive image patterns of each kind of shower. 

The reconstruction of the data being developed focuses on a system consisting of an image processing algorithm, related to knowledge-based classifiers, using only untreated (raw) images from Cherenkov telescopes. Such problem belong to the class of image categorization with a large number of entries, and are solved successfully by deep learning algorithms. Our group follows this approach , and currently the software TensorFlow is used for this purpose. It is also foreseen to use it for problems related to the reconstruction of the energy and determination of the original direction of the gamma ray.