The ALICE experiment at the LHC uses machine learning to identify subatomic particles emitted from heavy ion collisions. In 2017 ALICE is scheduled for an upgrade which will limit the bandwidth from the transition radiation detector to the particle identification algorithm to six 8-bit particle IDs (PID) per particle.
The aim is to investigate algorithms for the generation of PIDs and which machine learning algorithm is suited to processing them.
This project investigates the application of artificial neural networks rather than a summation approach for the generation of PID numbers for ALICE’s upgraded online algorithms.
This project investigates the applicability of boosted machine learning models rather than artificial neural networks for ALICE’s upgraded system.