Breakthrough in Artificial Intelligence: Researchers Develop Scalable Machine Learning Solution for ATLAS Detector

Strategic Insights -

Artificial intelligence has taken a significant leap forward with the development of a scalable machine learning solution for the ATLAS detector. Researchers from the University of Washington have created a tool called AthenaTriton that integrates the NVIDIA Triton Inference Server with the ATLAS software framework, Athena. This innovative solution enables the deployment of diverse machine learning models in the ATLAS environment, making it an essential step in the analysis of vast amounts of data from the ATLAS detector.

Key Takeaways:

  • AthenaTriton is a scalable machine learning solution developed by researchers at the University of Washington to address the growing demands of processing simulation and collision data in the ATLAS detector.
  • The tool integrates the NVIDIA Triton Inference Server with the ATLAS software framework, Athena, enabling the deployment of diverse machine learning models.
  • AthenaTriton operates as a client that sends requests to a remote or local server that performs the model inference, maximizing event throughput and utilizing coprocessors like graphics processing units (GPUs).
  • This scalable approach can be used in both online and offline computing workflows, meeting the challenges of processing vast amounts of data from the ATLAS detector.
  • Researchers have defined and implemented a dual-use machine learning interface with two distinct backends: the Open Neural Network Exchange (ONNX) Runtime and Inference as a Service.
  • The tool is being deployed in the ATLAS software framework to enhance the detector's analysis capabilities, including detector simulations, event reconstructions, and data analyses.
  • Chou Yuan-Tang and his research team at the University of Washington are leading the development of AthenaTriton.

Statistics:

  • The ATLAS detector collects vast amounts of data, requiring scalable machine learning solutions to process and analyze.
  • AthenaTriton can process event throughput of up to _________________ events per second, utilizing acceleration from graphics processing units (GPUs).
  • Researchers have demonstrated the capabilities of AthenaTriton in both online and offline computing workflows, showcasing its potential for real-time analysis.
  • The tool has been integrated into the ATLAS software framework, enhancing the detector's analysis capabilities in various domains, including detector simulations, event reconstructions, and data analyses.

Sources:

  • NewsRx. University of Washington Researchers Discuss Research in Machine Learning (AthenaTriton: A tool for running machine learning inference as a service in Athena). Journal of Engineering. October 27, 2025; p 4936.
  • Chou Yuan-Tang, et al. (2025). AthenaTriton: A tool for running machine learning inference as a service in Athena. EPJ Web of Conferences, 2025,337():01358. (EPJ Web of Conferences - http://www.epj-conferences.org/).
  • EPJ Web of Conferences. AthenaTriton: A tool for running machine learning inference as a service in Athena. (2025). EPJ Web of Conferences, 2025,337():01358. (EPJ Web of Conferences - http://www.epj-conferences.org/).