.jpg)
Researchers Jin Huang, Dmitrii Torbunov, Yihui "Ray" Ren, Yi Huang, and Yeonju Go stand in Brookhaven Lab's Scientific Computing and Data Facilities, where data from Relativistic Heavy Ion Collider collisions are stored.
Photo Credit: Timothy Kuhn/Brookhaven National Laboratory
Scientific Frontline: Extended "At a Glance" Summary: Unsupervised Artificial Intelligence in Particle Physics
The Core Concept: A generative artificial intelligence framework has been developed to autonomously isolate high-energy particle jet signals from the massive background noise generated during relativistic heavy-ion collisions. This unpaired, unsupervised model learns directly from simulated physics collision data to identify jets without relying on pre-labeled training examples.
Key Distinction/Mechanism: Unlike traditional supervised machine learning algorithms that require extensive human-annotated "answer keys" to train the model, this approach uses unsupervised learning to independently split crowded detector images into distinct signal and background components.
Origin/History: The research was led by the U.S. Department of Energy’s Brookhaven National Laboratory and published in Scientific Reports in 2026, utilizing data frameworks from the Relativistic Heavy Ion Collider (RHIC), which concluded operations in early 2026.
Major Frameworks/Components:
- Relativistic Heavy-Ion Collisions: The acceleration and collision of heavy ions at near-light speeds to recreate early-universe conditions.
- Quark-Gluon Plasma (QGP): A hot, dense state of matter composed of free quarks and gluons that existed microseconds after the Big Bang.
- Particle Jets: Collimated sprays of high-energy quarks and gluons used to probe the structural changes and properties of the QGP.
- Unsupervised Generative Learning: Artificial intelligence models that independently detect underlying patterns within massive data sets without explicit human categorization.
Branches of Science: High-Energy Physics, Nuclear Physics, Computational Physics, Particle Physics
Future Application: The autonomous algorithm can be applied to vastly improve data filtering across future high-energy collider experiments, minimizing the need for manual data annotation and accelerating the analysis of dense particle interactions.
Why It Matters: By improving background subtraction, this artificial intelligence method enables far more precise measurements of how quarks and gluons behave and interact, ultimately helping physicists decode the strong nuclear force and the fundamental properties of matter in the early universe.
.jpg) |
| This schematic shows jet production and detection in heavy-ion collisions. Red lines show heavy-ion beams. Yellow lines represent charged particle tracks. Blue boxes show calorimeters measuring energy. Light-gray cones show jet signals, which are difficult to identify due to the large underlying event background in these collisions. Image Credit: Brookhaven National Laboratory |
A team of physicists and computer scientists led by researchers at the U.S. Department of Energy’s (DOE) Brookhaven National Laboratory developed a new artificial intelligence (AI) approach that learns on its own to separate important high-energy particle jet signals from the enormous background of particles recorded in relativistic heavy-ion collisions, where two beams of heavy ions collide at near the speed of light. This methodological advancement isolates high-energy signals from background noise, enabling more precise measurements of the quark-gluon plasma (QGP), a hot, dense state of matter that existed a microsecond after the Big Bang.
At Brookhaven Lab, physicists recreated early-universe conditions that produce the QGP, a "soup" of free quarks and gluons, in heavy-ion collisions at the Relativistic Heavy Ion Collider (RHIC), a DOE Office of Science user facility for nuclear physics research that completed operations in early 2026. Physicists use jets—collimated sprays of high-energy particles produced in relativistic heavy-ion collisions—as one of the best probes of the QGP to glean more about how quarks and gluons behave and about the strongest force in nature that binds them together.
High-energy collisions produce energetic quark and gluon streams that are called jets. As jets stream through the QGP, the many particles within them lose energy and experience changes to their internal structure. Scientists want to zero in on those changes to learn more about the properties of the QGP.
A primary limitation in studying jets is that separating signal from the significant background produced in heavy-ion collisions has restricted measurement precision. The team's new AI approach tackles that challenge by learning to distinguish jet signals from the surrounding background more effectively than conventional methods, without relying on prelabeled examples.
"Improving background subtraction can enable more precise measurements of how quarks and gluons interact under extreme conditions, helping scientists better understand the strong force and the behavior of matter in the early universe," said Yeonju Go, a Brookhaven Lab physicist and Stony Brook University assistant professor working on RHIC data analysis.
In a paper recently published in Scientific Reports, the team, which included researchers from Columbia University and the University of Colorado Boulder, describes a generative AI method that separates the signal of jets from collision background—without requiring paired signal-and-background examples for training the model. This work shows the first use of an "unpaired, unsupervised" model that learned directly from simulated physics collision data, rather than from "this or that" labels fed to it by scientists.
"The detector sees a crowded ‘image’ in which the signal we care about is buried inside a dense spray of other particles," Go said. "Our AI method learns to split that image into two parts—the jet signal and the background—without requiring one-to-one 'answer key' examples that are not available in real experimental data."
.jpg)
This diagram shows how the framework is trained and then used to separate a particle jet's signal from the surrounding background in collision data.
Image Credit: Brookhaven National Laboratory
A New Approach
The AI framework’s ability to learn directly from data sets it apart from conventional methods, researchers said.
"The problem with the traditional approach is that it requires a lot of human work to annotate all the data," Brookhaven Lab computer scientist Dmitrii Torbunov said. "The newer generation of machine learning algorithms focus on the so-called unsupervised learning paradigm."
Traditional supervised learning algorithms require researchers to manually annotate extensive training datasets, labeling thousands of collision events to teach the model to identify specific particle signatures. The new model doesn’t require all that labeling, Torbunov said.
"Unsupervised learning methods can learn useful concepts and patterns directly from the data, without requiring humans to label every example," he said.
The research team tested the physics-optimized model using simulated data captured by particle detector components called calorimeters at sPHENIX, a detector that operated during RHIC’s final three runs. In real sPHENIX data, jet signals are immersed in an extremely dense and fluctuating heavy-ion collision environment.
Researchers provided the model with simulated detector data that included collision background, along with separate simulated jet images without background.
"What our algorithm does is look at a collection of pure background data produced by the QGP and at our ideal jet simulations, and then automatically figure out how to extract the underlying jet signal while removing the background," Torbunov said. "This new paradigm is extremely useful for science because scientific data can contain very large backgrounds, and from a human perspective it may be difficult, or even impossible, to manually determine how much of each data point comes from the background and how much comes from the jet signal."
In simulated sPHENIX data, the model more accurately reconstructed jets' positions, energies, and internal structure than conventional background-subtraction methods, particularly for larger jets where background fluctuations are strongest.
Looking ahead, researchers aim to apply the model to the wealth of physics data collected by sPHENIX and possibly apply it to future experiments at the Electron-Ion Collider (EIC). During its operational runs, sPHENIX collected nearly 300 petabytes of experimental data.
"We definitely want to use the model on our experimental data, especially as sPHENIX has a very large data set," Go said. "So that’s the next step: extracting the physics results out of it. This would be the first use of this kind of model in the nuclear and particle physics community, not just RHIC, and we expect it to be applicable to calorimeter-based measurements at the EIC or the Large Hadron Collider in Europe, and other experiments where separating signal from background is a fundamental challenge."
Because the framework works directly with image-like detector data, researchers say its potential extends beyond high-energy particle jets studies to other experiments that must disentangle overlapping signals.
Cross-Lab Collaboration
The new machine learning approach comes amid a growing focus across DOE National Laboratories and other research institutions to harness AI to accelerate scientific discovery. This work is one additional example of how researchers at Brookhaven, a multidisciplinary laboratory, are collaborating to advance AI for science.
"The integration of computer science and physics at Brookhaven Lab was critical to developing this machine learning model," Go said. "This is not a one-time project, and we expect further opportunities to integrate AI and machine learning in our scientific operations."
This work also ties to a DOE Genesis Mission Phase I award, led by Baruch College, that Brookhaven Lab is contributing to along with Stony Brook University, Columbia University, and the University of Colorado Boulder: Characterizing Jet Modification in the Quark-Gluon Plasma Using Unsupervised, Cycle-Consistent Generative Learning. Go and Torbunov each play leading roles in the project.
Funding: This research was supported by Brookhaven’s Laboratory Directed Research and Development funding and the DOE Office of Science.
Published in journal: Scientific Reports
Authors: Yeonju Go, Dmitrii Torbunov, Yi Huang, Shuhang Li, Timothy Rinn, Haiwang Yu, Brett Viren, Meifeng Lin, Yihui Ren, Dennis Perepelitsa, and Jin Huang
Source/Credit: Brookhaven National Laboratory
Edited by: Scientific Frontline
Reference Number: phy111326_01
