Ultra-fast deep learning algorithms on FPGA for the phase-II level-0 trigger of the ATLAS experiment

The LHC accelerator will face, during the following years, a complete upgrade with the main purpose of rising up the instantaneous luminosity by a factor of almost five. Though this will permit to collect an incredible amount of data, the complexity of each event will greatly intensifies going from an average number of interactions per bunch crossing of 40 to an average of 200. To cope with this problem and be able to handle this large amount of information, both the detectors and the trigger algorithms of the ATLAS experiment will be updated. A machine learning approach for the level-0 trigger algorithm is presented.