The ATLAS and CMS collaborations are using cutting-edge machine learning techniques to search for exotic-looking collisions that could indicate new physics. Credit: S Sioni/CMS-PHO-EVENTS-2021-004-2/M Rayner
Artificial intelligence is revolutionizing the way LHC experiments detect new particles.
By training AI to recognize and distinguish between normal and anomalous jets, researchers can identify potential new physics hidden within particle collisions. A physics conference highlighted recent developments, demonstrating the growth and potential of these AI applications.
One of the main goals of the Large Hadron Collider (LHC) experiments is to find signatures of new particles, which could explain many of the unsolved mysteries of physics. Often, searches for new physics are designed to find one specific type of new particle at a time, using theoretical predictions as a guide. But what about finding unexpected – and unexpected – new particles?
The billions of collisions in the LHC's experiments would be a daunting task for physicists without knowing exactly what to look for. So, instead of digging through the data and finding anomalies, ATLAS and CMS collaboration is enabling Artificial intelligence (AI) Streamline the process.
AI advances in particle detection
I Rencontres de Moriond At the conference on March 26, physicists in collaboration with CMS presented the latest results obtained using different methods. Machine learning Techniques for finding pairs of “jets”. These jets are collimated sprays of particles produced by strongly interacting quarks and gluons. They are particularly difficult to analyze, but they can hide new physics.
The event display of one of the CMS events determined by the AI algorithm is highly irregular and therefore likely coming from a new particle. Credit: CMS Collaboration
Techniques in AI Training for Physics
ATLAS and CMS researchers use several strategies to train AI algorithms to search for jet planes. By studying the shape of their complex energy signatures, scientists can determine which particle created the jet. Using data from real collisions, physicists in both experiments are training their AI to recognize the characteristics of jets produced by known particles. The AI is then able to distinguish between these jets and unusual jet signatures, which potentially indicate new interactions. These will appear as accumulations of atypical jets in the data set.
Another method involves instructing the AI algorithm to consider the entire collision event and detect unusual features in the detection of different particles. These unusual features may indicate the presence of new particles. The technique was demonstrated in a paper released by ATLAS in July 2023, featuring one of the first applications of unsupervised machine learning resulting from the LHC. In CMS, in a different approach physicists create simulations of possible new signals and then task AI with identifying collisions in real data that differ from regular jets but resemble the simulations.
The impact of machine learning on particle physics
In the latest results presented by CMS, each AI training method showed different sensitivity to different types of new particles, and no single algorithm proved to be the best. The CMS team was able to constrain the production rate of several different types of particles that produce the unusual jet. They were also able to show that the AI-led algorithm significantly increased the sensitivity of a wide range of particle signatures compared to traditional techniques.
These results show how machine learning is revolutionizing the search for new physics. “We already have ideas on how to further improve the algorithm and apply it to different parts of the data to find many types of particles,” says Oz Amram from the CMS analysis team. How to apply.”