UM-Bridge: Bridging the Gap Between UQ and Model Software

Linus Seelinger (U Heidelberg)

May 03. 2022, 10:00 — 10:30

Advanced UQ algorithms are not frequently coupled with advanced model software. While the link between UQ and models is often conceptually simple, implementations are challenging: Incompatible programming languages or software frameworks might be used in the two communities, software stacks may become exceedingly complex, and experts from both sides are required.

UM-Bridge addresses those issues by providing an abstract and easy to use software interface for linking UQ and models across programming languages. By leveraging containers, portability of complex models and separation of concerns between developers can be achieved.

Finally, UM-Bridge allows the definition of reproducible, near-black box UQ benchmark problems that can be accessed from any UQ framework. The main discussion point is what specific problems should be included in a standardized UQ benchmark suite, and we hope to get the wider community involved in their definition and implementation.

Further Information
ESI Boltzmann Lecture Hall
Associated Event:
Computational Uncertainty Quantification: Mathematical Foundations, Methodology & Data (Thematic Programme)
Clemens Heitzinger (TU Vienna)
Fabio Nobile (EPFL Lausanne)
Robert Scheichl (U Heidelberg)
Christoph Schwab (ETH Zürich)
Sara van de Geer (ETH Zürich)
Karen Willcox (U of Texas, Austin)