Democratizing Uncertainty Quantification

TL;DR

UM-Bridge introduces a universal, language-agnostic protocol that connects uncertainty quantification (UQ) software with forward simulation models, enabling seamless separation of concerns and scalable computation on HPC and cloud resources. The framework is demonstrated through a substantial benchmark library and multiple real-world applications, including naval hull resistance, composite material defects, and tsunami source inversion, all facilitated by containerization and flexible parallelization. By decoupling UQ, models, and HPC infrastructure, UM-Bridge enables reproducible benchmarking across diverse methodologies and languages, fostering rapid development and cross-domain collaboration. The work provides a practical pathway to democratize UQ, offering standardized interfaces, portable models, and scalable execution that can adapt to growing computational demands in science and engineering.

Abstract

Uncertainty Quantification (UQ) is vital to safety-critical model-based analyses, but the widespread adoption of sophisticated UQ methods is limited by technical complexity. In this paper, we introduce UM-Bridge (the UQ and Modeling Bridge), a high-level abstraction and software protocol that facilitates universal interoperability of UQ software with simulation codes. It breaks down the technical complexity of advanced UQ applications and enables separation of concerns between experts. UM-Bridge democratizes UQ by allowing effective interdisciplinary collaboration, accelerating the development of advanced UQ methods, and making it easy to perform UQ analyses from prototype to High Performance Computing (HPC) scale. In addition, we present a library of ready-to-run UQ benchmark problems, all easily accessible through UM-Bridge. These benchmarks support UQ methodology research, enabling reproducible performance comparisons. We demonstrate UM-Bridge with several scientific applications, harnessing HPC resources even using UQ codes not designed with HPC support.

Figures (19)

• Figure 1: UM-Bridge connecting UQ and numerical models through a universal interface.
• Figure 2: Fortran code for modelling the pressure distribution and the wave elevation generated by a vessel. UM-Bridge support is achieved through a simple wrapper, enabling parallelized from, e.g., Matlab (left figure). FE model of a carbon fibre composite aerospace component implemented in a complex solver in C++. Through UM-Bridge and containerization, a expert could solve uncertainty propagation without technical knowledge of the model code (center & right figures).
• Figure 3: Monolithic coupling of and model software in a single application (left); UM-Bridge providing a universal interface between and model applications (right). UM-Bridge integrations ("UM") handle network communication behind the scenes. Optionally, models may be containerized.
• Figure 4: UM-Bridge provides a general-purpose Kubernetes setup for scaling up applications. It runs many parallel instances of containerized models and distributes concurrent requests from a client among them. Any UM-Bridge supporting or model code readily works with this setup.
• Figure 5: Synthetic weak scaling test of the kubernetes setup on for various numbers of parallel model instances, requesting a constant number of evaluations per model instance.