Deep learning at the LHC

Markus Stoye
(Imperial College)

In the last few years deep learning revolutionized the fields of computer vision, computer speech recognition, and other tasks. This has laid the foundation for self-driving cars and many other exciting applications. One pillar of this successes was the huge amount of available data. Clearly, the amount of data at the LHC is of outstanding scale. After a brief introduction to deep learning, I present real life examples of deep learning applications in collider experiments and give an outlook on current trends, such as addressing real world data and simulation differences, opening the “black box”, as which deep neural networks are perceived, and using prior knowledge from matrix element calculation for training.