[ONLINE] Machine learning prototype for SPACE applications

Annotation

The objective of this webinar is to present the prototype of a machine learning tool to enable the exploration, analysis, and interpretation of the outputs of large-volume cosmological simulations using Representation Learning techniques. The tool efficiently learns a low-dimensional representation of the structure of simulated galaxies in arbitrary physical components, uncovering their intrinsic structural distribution. It also provides an interactive hierarchical visualization of the entire simulation and its compact representation, and scales to arbitrarily large simulations beyond the Exascale era.

Language

English

Level

beginner 

Tutors:

Sebastian Trujillo Gomez (HITS) and Bernd Doser (HITS) 

Sebastian Trujillo Gomez is currently a researcher in the Astroinformatics group at the Heidelberg Institute for Theoretical Studies, where he develops machine learning methods to drive scientific discovery using complex numerical simulations. He holds a Ph.D. in Astrophysics and has spent most of his career trying to understand how galaxies formed after the Big Bang, and how to use them as probes of cosmology and the elusive nature of dark matter. Before joining HITS in 2023, he worked as a postdoc at the University of Zurich and at Heidelberg University.

Bernd Doser is currently employed as a Senior Scientific Software Engineer at the Heidelberg Institute for Theoretical Studies gGmbH (HITS). His role involves implementing modern software engineering methods, and contributing to and maintaining several open-source packages for various scientific disciplines. Before joining HITS, he worked at Avant-Garde Materials Simulation GmbH, where he was responsible for crystal structure prediction and force field generation. Bernd holds a Ph.D. in computational chemistry and is an expert in high-performance computing, C++, and machine learning.

Acknowledgements

This work was supported by the SPACE project under grant agreement No 101093441. The project is supported by the European High-Performance Computing Joint Undertaking and its members (including top-up funding by the Ministry of Education of the Czech Republic ID: MC2304).

This course was supported by the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID:90254).