A pioneering algorithm has been put into use at the Large Hadron Collider by a group of scientists from Staffordshire University, Massachusetts Institute of Technology, and CERN.
The Large Hadron Collider (LHC), the most potent particle accelerator ever created, is located at CERN, the European Organization for Nuclear Research, close to Geneva in Switzerland. It is housed in a tunnel 100 meters below ground. It is where experiments are done that help physicists from all over the world learn more about the universe.
The undertaking is a component of the Compact Muon Solenoid (CMS) experiment, one of seven operational experiments that use detectors to examine the particles generated by collisions in the accelerator.
The project was completed in advance of the Large Hadron Collider’s high luminosity upgrade and is the subject of a recent academic paper titled End-to-end multiple-particle reconstruction in high occupancy imaging calorimeters with graph neural networks. The High Luminosity Large Hadron Collider (HL-LHC) project seeks to boost the LHC’s efficiency in order to increase the likelihood of scientific breakthroughs after 2029. Proton-proton interactions will go from 40 to 200 during an event thanks to the HL-LHC.
The study’s principal investigator is Professor Raheel Nawaz, Pro Vice-Chancellor for Digital Transformation at Staffordshire University. He said: “We are advocating the use of modern machine learning techniques to perform particle reconstruction as a possible solution to this problem. Limiting the increase of computing resource consumption at large pileups is a necessary step for the success of the HL-LHC physics program. ”
Using a more advanced AI-based solution, he continued, “This project has been both a joy and a privilege to work on, and it is likely to determine the future direction of particle reconstruction research.”
Additionally, Dr. Jan Kieseler from CERN’s Experimental Physics Department said: “This is the first single-shot reconstruction of a proton-proton collision with about 1000 particles from and in a previously difficult environment with 200 simultaneous interactions. An important step toward future particle reconstruction is demonstrating that this innovative method, which combines specialized graph neural network layers (GravNet) and training techniques (Object Condensation), can be extended to such difficult tasks while adhering to resource constraints.
