Abstract
The recent boom in artificial intelligence (AI) alongside big data
being collected everywhere in our world, is impacting our lives in
an unprecedented way. The growing exploration of machine learning
algorithms in particle physics offer new solutions in many areas
such as analyses, simulation and reconstruction. However, the large
number of computations also make it extremely challenging to program
these algorithms on more energy-efficient, faster computing hardware
such as field programming gate arrays (FPGAs) or
application-specific customized chips (ASIC), thus restricts the
usage of these networks in applications that have stringent speed
requirements, such as in large-scale particle physics experiments,
which need to handle large data volume at a high rate. Accelerated
machine learning inference can potentially provide solutions to
these challenging demands for both high-performance trigger as well
as high-throughput computing systems. In my talk, I will discuss
developments in employing accelerated machine learning inferences as
solutions to our data processing challenges in particle physics.