To path indicator Subpages of “University“ To navigation by target groups To navigation by topic To quick access To footer with other services
To content
Sunset at the North Campus © Roland Baege​/​TU Dortmund
PROJECT FUNDED BY THE FEDERAL MINISTRY OF EDUCATION AND RESEARCH

New Algorithms for Particle Physics

-
in
  • Particle Physics
  • Top News
  • Research
The image shows the simulation of a particle collision in the ATLAS experiment. © ATLAS
The picture shows the simulation of a particle collision in the ATLAS Experiment.

The Federal Ministry of Education and Research will fund a new interdisciplinary collaborative project over the next three years, in which physicists from TU Dortmund University are also participating. In the project entitled “Artificial Intelligence for Rapid Simulation of Scientific Data” (KISS), researchers are developing AI-based simulation methods. In the future, this should make it possible to analyze the vast amounts of data generated in particle and astroparticle physics more quickly and efficiently. KISS is being funded within the ministry’s “ErUM-Data Action Plan”. Of the total funding of about €3.5m, around €500,000 will go to TU Dortmund University.

Already today, particle detectors and telescopes generate vast amounts of data and analyzing these data necessitates immense computing resources. As such detectors and telescopes become even more advanced, computing capacities will soon reach their limits. “That is why we have to work on faster and more efficient algorithms in order to use the available resources in a sustainable way,” explains Professor Kevin Kröninger, who is in charge of a subproject within the collaboration. To this end, researchers from particle and astrophysics from throughout the whole of Germany are developing new AI-based simulation processes within the KISS project. Teams from TU Dortmund University led by physics professors Kevin Kröninger and Wolfgang Rhode are participating in the project, which is coordinated by the University of Hamburg.

Professor Kevin Kröninger and his team are conducting research as part of the ATLAS experiment at the Large Hadron Collider (LHC), the most powerful particle accelerator in the world. In their research, the scientists work a lot with simulations. “The problem is that the simulation methods used today are slow and often a limiting factor when it comes to application,” Kröninger explains. This is where his research in the KISS project comes in: He is concentrating on a special class of algorithms – known as Markov chain Monte Carlo methods (MCMC) – that are to be used in simulations in the future. The aim is to use machine learning here as well. “These innovative algorithms can considerably reduce computing time and ultimately lead to better comparisons between measurements from experiments and predictions through simulations,” says Kröninger. In his subproject, he is working closely with Professor Katja Ickstadt from the Department of Statistics.

Professor Wolfgang Rhode’s group is conducting research in a second subproject. His team is working on large-scale experiments in astroparticle physics, such as the IceCube Neutrino Observatory at the South Pole or MAGIC, the high-energy gamma-ray telescope on La Palma. The telescopes detect energy-rich particles from space, and here, too, the physicists need simulations to analyze the data streams. Within the KISS project, the team led by Professor Rhode wants to develop a system that facilitates smart decision-making: Is it necessary to produce new simulations for observations? Or are there cases where it is possible to fall back on existing simulation datasets? Here, the physicists are cooperating closely with scientists from the Lamarr Institute for Machine Learning and Artificial Intelligence.

About the funding: With its ErUM-Data Action Plan, Germany’s Federal Ministry of Education and Research supports projects to master the challenges of digitization in basic scientific research at large-scale facilities.

Contact for inqueries: 

Back

To top of page