Advanced Data Analysis Methods for Experiments at Jefferson Lab and the Electron-Ion Collider
ORAL · Invited
Abstract
Jefferson Lab, located in Newport News, Virginia, hosts a leading nuclear physics research program dedicated to understanding the fundamental building blocks of matter. The presentation will explore how researchers are using modern data analysis tools and techniques to solve complex scientific problems, with examples that illustrate both the challenges and the innovative solutions being developed. The role of AI and data science in transforming nuclear physics data analysis will be discussed throughout the presentation.
The key challenge is to distill information about a hidden quantum world involving a large and varying number of particles—quarks and gluons—that are all relativistic and strongly interacting, from data collected by detecting other particles. Due to the quantum nature of the world at the smallest scales, nothing is definite. Because of the strong interactions among the particles, data analysis from experiments involves many overlapping possible states.
For example, scientists are searching for new forms of matter, such as exotic particles made from gluons, using advanced methods that account for hundreds of possible outcomes. These approaches must be carefully designed to produce results that are accurate and reproducible. Another example is the extraction of information about the underlying three-dimensional images of the distributions of quarks and gluons and their interactions inside protons and neutrons from measured data that are incomplete and imprecise. This requires the collection and analysis of very large, detailed data sets that involve multiple variables and final particle states.
This challenging quantum environment has driven—and continues to drive—the development of powerful analysis tools and methods that enhance scientific discovery and improve precision. Building on these achievements, the talk will explain how Jefferson Lab’s expertise is now being leveraged in the design and planning of data analysis efforts for the upcoming Electron-Ion Collider (EIC). This discussion will focus on the ePIC experiment at the EIC and will present both the development and implementation of a computing model designed to accelerate analysis workflows, as well as efforts to bridge experiment and theory—enabling joint progress on both fronts.
The key challenge is to distill information about a hidden quantum world involving a large and varying number of particles—quarks and gluons—that are all relativistic and strongly interacting, from data collected by detecting other particles. Due to the quantum nature of the world at the smallest scales, nothing is definite. Because of the strong interactions among the particles, data analysis from experiments involves many overlapping possible states.
For example, scientists are searching for new forms of matter, such as exotic particles made from gluons, using advanced methods that account for hundreds of possible outcomes. These approaches must be carefully designed to produce results that are accurate and reproducible. Another example is the extraction of information about the underlying three-dimensional images of the distributions of quarks and gluons and their interactions inside protons and neutrons from measured data that are incomplete and imprecise. This requires the collection and analysis of very large, detailed data sets that involve multiple variables and final particle states.
This challenging quantum environment has driven—and continues to drive—the development of powerful analysis tools and methods that enhance scientific discovery and improve precision. Building on these achievements, the talk will explain how Jefferson Lab’s expertise is now being leveraged in the design and planning of data analysis efforts for the upcoming Electron-Ion Collider (EIC). This discussion will focus on the ePIC experiment at the EIC and will present both the development and implementation of a computing model designed to accelerate analysis workflows, as well as efforts to bridge experiment and theory—enabling joint progress on both fronts.
*The speaker is supported by the U.S. Department of Energy Office of Science, Office of Nuclear Physics contract number DE-AC05-06OR23177, under which Jefferson Science Associates, LLC operates Jefferson Lab.
–
Presenters
-
Markus Diefenthaler
- Jlab