Using Deep Learning to observe the Higgs Boson in Real-time

• Philip Harris (MIT)

Abstract: With a raw data rate exceeding 1 Petabit per second, particle detectors at the Large Hadron Collider(LHC) at the Europe Center for Nuclear Research (CERN) contend with some of the largest data rates ever encountered. With planned upgrades in the near future, these rates will continue to grow, further complicating our ability to process data effectively to continue understanding the universe's fundamental properties. To process data in real-time, we rely on specialized computing systems using Field Programmable Gate Arrays (FPGAs) and, in later stages of data acquisition, Graphics Processor Units(GPUs). Despite the enormous challenge, we are finding that deep learning is not only capable of handling these incredible data rates but is significantly enhancing our ability to process data in real-time. In this talk, we present a strategy for integrating deep learning into the LHC data acquisition. We show how it dramatically improves the quality of physics measurements we can perform downstream. Finally, we discuss the scientific implications emerging from this work in a broad range of scientific fields, including nuclear physics, astrophysics, materials science, and neuroscience.