Kyungmin Park Accelerates CMS Trigger Work at CERN
Heidi Opdyke
- Associate Dean of Marketing and Communications, MCS
- Email opdyke@andrew.cmu.edu
- Phone 412-268-9982
Carnegie Mellon University doctoral student Kyungmin Park is helping scientists make faster, smarter decisions about which particle collisions to study at CERN, the world’s largest particle physics laboratory located near Geneva. Her work is part of the Next Generation Trigger (NextGen) project, which aims to revolutionize how data is filtered and analyzed at the Large Hadron Collider (LHC).
Park, a fifth-year Ph.D. candidate in CMU’s Department of Physics works with her advisor Professor Manfred Paulini on the Compact Muon Solenoid (CMS) experiment. The CMS is one of the major detectors at the LHC where protons collide at nearly the speed of light — creating conditions similar to those just after the Big Bang.
These collisions happen at the rate of about one billion every second. Scientists can’t store or analyze all the data. That’s where the trigger system comes in — a kind of ultra-fast decision-maker that filters out the noise and keeps only the most promising events.
“The trigger needs to decide on whether a collision is interesting or not and quickly reconstructs very vaguely what happened in the particle interactions,” Park said.
Park is working to improve how the trigger identifies important events using artificial intelligence, specifically by designing algorithms that run directly on hardware. The algorithms must be tiny, lightning fast and incredibly efficient because they operate in real time as collisions happen.
The CMS experiment uses two levels of triggers:
- Level-1: Makes decisions in less than 10 microseconds — 1/1,000 the blink of an eye. It runs in sync with particle bunches colliding in the detectors and uses a combi-nation of algorithms that save information on collisions with characteristics of interest.
- High-Level Trigger: Further refines the data, ultimately saving up to 1,000 events per second for deeper analysis.
Park began her work on machine learning algorithms related to the Level-1 Trigger under CMU Assistant Professor Matteo Cremonesi. Her early success impressed her mentors.
“She executed her first task so perfectly and in such a short time that Professor Paulini and I wondered how we could challenge her a bit more,” Cremonesi said. “Kyungmin is by far one of the most brilliant students at CERN currently working on the trigger and machine learning in real time.”
Park, who has undergraduate degrees in physics with a minor in computer science from the University of Seoul, has always been interested in AI applications in physics. Her talent was recognized during a week-long trigger hackathon, where she caught the attention of CERN senior scientist Cristina Botta. Botta encouraged Park to apply for the NextGen fellowship, supported by the Eric and Wendy Schmidt Foundation. The fellowship allows Park to dedicate half her time to developing future NextGen projects under the supervision of Botta and other CERN physicists.
“It’s some of the most fun stuff. Every decision about the collision needs to happen quickly in the hardware, and collisions happen very frequently, so the AI algorithms must be small, efficient and fast enough,” Park said. “It’s an interesting problem to think about. Changing the very first step can impact the entire chain of events. Whatever you lose at this very first step is gone and does not get to be analyzed. And at the end, it impacts the physics we can study with CMS.”
Paulini emphasized the broader impact of her skills.
“Kyungmin is an exceptionally multi-talented student who fully embraces the opportunities of working at the forefront of computing, developing state-of-the-art AI algorithms, improving advanced instrumentation, communicating effectively, presenting impactful results and cultivating a broad set of foundational skills that students pick up during their Ph.D. in particle physics,” Paulini said. “These are the skills that are directly transferable to solving real-world challenges and advancing their future career.”
Park has been at CERN for more than two years and will continue her work there through the end of her Ph.D. She chose Carnegie Mellon not only for its strong connections to the CMS experiment but also for of its interdisciplinary nature and caring environment.
While in Pittsburgh, she served as the Department of Physics outreach coordinator and played drums in several music projects.
“At CMU, there’s traditionally a lot of art meeting science, and even if art is not directly related to my graduate studies, I do music all the time. It was a very interesting atmosphere that I wanted to join,” she said. “And CMU's Physics Department provides a very supportive environment.”
