Digital modelling has long been central to scientific and engineering practice, providing predictive tools for analysing complex systems. Traditionally, such models have been employed in a forward manner, where known parameters yield simulated outcomes. The complementary task – inferring unknown pa-rameters or latent structures from observed data – has historically been more challenging. Recent advances in optimization and statistical inference, not least those developed in the context of deep learning and generally in the field artificial intelligence, have significantly expanded the methodological repertoire available for such inverse problems. As a result, inverse simulation has matured into a robust framework with broad applicability across disciplines.
This methodological progress has enabled new classes of applications. Inverse simulation now underpins advances in areas as diverse as material characterization, biomedical imaging, and environmental modeling, where the ability to reconstruct hidden properties from indirect measurements is essential. The increasing ubiquity of such problems highlights the need for researchers who are not only proficient in the theoretical underpinnings of inverse methods but also capable of applying them effectively across heterogeneous domains.
The proposed Research Training Group (GRK) addresses this need by establishing a structured doctoral program for 30 PhD students. The program emphasizes a dual focus: (i) rigorous training in the mathematical and computational foundations of inverse simulation, including optimization, uncertainty quantification, deep learning, foundation models and model regularization; and (ii) exposure to a broad spectrum of application domains, ensuring that methodological advances are informed by real-world challenges.
By fostering interdisciplinary collaboration and combining methodological rigor with application-driven research, the GRK aims to advance inverse simulation as a unifying framework in computational science and technology. The program will contribute both to the development of novel algorithms and to the training of researchers who can bridge theory and practice, equipping them to address the inherent ambiguity and uncertainty of inverse problems in real-world applications.
Please note: This is just an information regarding events taking place at SSC; public attendance is therefore not possible due to the character of the retreat.
