Title:	Robust and scalable forward and inverse learning
DNr:	Berzelius-2025-60
Project Type:	LiU Berzelius
Principal Investigator:	Prashant Singh <prashant.singh@it.uu.se>
Affiliation:	Uppsala universitet
Duration:	2025-02-14 – 2025-09-01
Classification:	10210
Homepage:	https://www.prashantsingh.se
Keywords:

Abstract

Data-driven models are accelerating scientific workflows across disciplines. As a forward process, machine learning models are used for predictive modeling based on experimental data, or as fast surrogate models of complex simulators. Within inverse problems, machine learning models are used for parameter estimation and system analysis. The project explores robust and scalable deep learning approaches for both forward and inverse modeling problem settings. We consider applications in particle physics, microbiome studies and exposomics.