RODEM

Robust Deep Density Models for
High-Energy Physics and Solar Physics

High Energy Physics and Solar Physics face similar challenges.

Finding relevant cases in this data deluge

can potentially revolutionize physics!

But we need to find the needle in the haystack.

RODEM in a nutshell

RODEM is a SNSF/SINERGIA project fostering cooperation between high energy physicists, solar physicists and experts in machine learning in Switzerland to advance research methodologies in both fields.

During the last decade, the amount of data available to scientists has increased enormously. New infrastructures, such as the Large Hadron Collider (LHC), and a new generation of solar observatories, such as the Solar Dynamics Observatory (SDO), produce data on a scale that cannot be exploited to their full extent with existing methods.

Simultaneously, data science has experienced real game changing breakthroughs in the past years. In particular, deep learning methods have shown the potential of data driven approaches compared to traditional algorithmic approaches. The following questions will shape research stragegies.

➤  Can data driven methods support us in unraveling new physics?

➤  Can physics support us in making better deep learning models?

Now is the time to unite our experience in both physics and machine learning to deeply dive into the many challenges of this combination. The rewards will undoubtedly bring both domains forward, resulting in better forecasting tools, generative models and anomaly detectors to be applied in the fields of High Energy Physics and Solar Physics.

First results – conference papers
 

• Information-theoretic stochastic contrastive conditional GAN: InfoSCC-GAN, Bayesian Deep Learning NeurIPS 2021 Workshop [2112.09653] (2021)
• Funnels: Exact maximum likelihood with dimensionality reduction, Bayesian Deep Learning NeurIPS 2021 Workshop [2112.08069] (2021)
• Generation of data on discontinuous manifolds via continuous stochastic non-invertible networks, Bayesian Deep Learning NeurIPS 2021 Workshop [2112.09646] (2021)
• Turbo-Sim: a generalised generative model with a physical latent space, NeurIPS Workshop: Machine Learning and the Physical Sciences [2112.10629] (2021)