ATLAS Uses AI for First Full-Phase-Space

The ATLAS collaboration has completed the first full-phase-space measurement of the Z+jets production cross section at the Large Hadron Collider. This work, detailed in a thesis submitted on 28 August 2026, employs novel artificial intelligence techniques to unfold data across a high-dimensional space.

Proton-proton collisions at the LHC generate vast, complex datasets. The high collision rates and numerous final-state particles challenge traditional analysis methods. This complexity necessitates advanced techniques to extract all available information about fundamental interactions.

The AI-Driven Measurement

The core achievement is a differential cross-section measurement that accounts for the kinematics of every charged particle produced alongside the Z boson. An AI-based unfolding algorithm makes this comprehensive analysis possible. This approach provides a complete experimental characterization of the Z+jets process, a common signature in proton collisions.

The measurement serves as a proof-of-principle. It demonstrates the feasibility of applying such high-dimensional analyses to a host of other LHC processes. The technique illuminates known interactions and enhances the search for new physics hidden within complex event topologies.

Significance for High-Energy Physics

This work represents a significant methodological advance. Prior measurements often integrated over many kinematic variables or focused on specific, restricted regions. The AI-driven method avoids such compromises, exploiting the full information content of the ATLAS detector's recordings.

The complete characterization of Z+jets production offers stringent tests for theoretical predictions. Precise calculations of quantum chromodynamics and electroweak processes can now be compared against data in unprecedented detail. Discrepancies could point towards physics beyond the Standard Model.

Researchers note that the datasets from detectors like ATLAS are inherently high-dimensional. Novel analysis techniques are not merely beneficial but essential. Artificial intelligence provides the tools to navigate this complexity, transforming raw collision data into precise physical insights.

The thesis, titled "A High- and Variable-Dimensional Measurement of the Z+jets Differential Cross Section with the ATLAS Experiment and Artificial Intelligence," is available on the arXiv preprint server. It centers on applications of AI to improve the physics results of the ATLAS experiment. The successful implementation marks a step toward more exhaustive use of LHC data in the years to come.