workshop

UM-Bridge Workshop 2025

UM-Bridge logo

UM-Bridge is a language-agnostic interface linking Uncertainty Quantification (UQ) packages and numerical model software. It allows

For more information, see UM-Bridge documentation.

This workshop has the following goals:

Organizers

Time and Location

The workshop will be held online and span two half days:

Registration

Registration is available through the registration form.

Prerequisites

For practical exercises, you will need working installations of

Schedule

Tuesday, 9. December 2025

Time Session
13:00 Welcome
13:05 Anne Reinarz (Durham University): Introduction to UM-Bridge
13:40 Practical: UM-Bridge basics / Participants’ projects
14:20 Anne Reinarz (Durham University): Intro containers, containerized UM-Bridge models/benchmarks
14:40 Practical: Containerized models / Participants’ projects
15:00 Coffee break
15:30 Chung Ming Loi (Durham Uni.): UM-Bridge models in the cloud and on classical HPC
16:00 Practical: UM-Bridge cloud + HPC / Participants’ projects
17:00 Wrap up

Wednesday, 10. December 2025

Time Session
13:00 Invited talk: William Hornsby, UKAEA
13:30 Louise Kluge: Bayesian Inference in Expensive Models using MLDA and GP-Surrogates
  Vikas Kurapati: Fused ensembles of dynamic‑rupture earthquake simulations to accelerate Bayesian inference
14:00 Short break / discussion
14:30 Invited talk: Amal Mohammed A Alghamdi, Technical University of Denmark
15:00 Lightning talks (various)
15:30 Coffee break
16:00 Short talks
16:20 Invited Talk: Linus Seelinger, Pasteur Labs
16:50 Closing Remarks / Wrap-up

Abstracts

Amal Mohammed A Alghamdi: Bayesian inversion of CT data to characterize transport in the mouse ear

Recent studies in rodents and humans show that gene therapy agents or tracers injected into the cerebrospinal fluid (CSF) reach the inner ear. The communication of fluid between the cochlear and the subarachnoid spaces of the brain has been controversial for decades. Observations of transport between the cochlea and the subarachnoid space contrast with the different composition of the fluids. The recent discovery of a membrane in the cochlear aqueduct raises further questions about the restrictions of transport between the compartments. This study aims to numerically quantify the diffusive and advective modes of transport of inert molecules from CSF to an intact cochlea. We use imaging data of the transport of a small tracer through the cochlear aqueduct in five sedated mice (8-week-old males). To estimate the transport model parameters, we formulate and solve a Bayesian inverse problem in which we allow the diffusivity to vary in the presence of potential membranes. We also discuss how modeling choices affect the inference. We carry out the implementation using the software tool CUQIpy (Computational Uncertainty Quantification for Inverse Problems in Python). In this talk we also introduce CUQIpy-UMBridge plugin which allows CUQIpy models and distributions to be served via an Um-Bridge server.

William Hornsby: The use of turbulence surrogate models in plasma integrated modelling

Abstract: Plasma micro-turbulence is one of the dominant transport mechanisms of heat from the core of a fusion power plant. Direct numerical calculation of the micro-instabilities that form turbulence is computationally expensive and is a significant bottleneck in integrated plasma modelling, in which the many physical processes are coupled to predict reactor-level behaviour and to optimise operational scenarios of fusion power plants. The considerable number of geometric and thermodynamic parameters, the interactions that influence the turbulence and the resolutions needed to accurately resolve these turbulent modes, makes direct numerical simulation for parameter space exploration computationally extremely challenging. However, this makes it suitable for surrogate modelling, where speed ups of up to 105 are possible making rapid scenario development a possibility. In this talk the integrated plasma modelling use-case will be introduced as well as the turbulence surrogate modelling efforts at UKAEA, including how the models are integrated into larger workflows.

Vikas Kurapati: Fused ensembles of dynamic‑rupture earthquake simulations to accelerate Bayesian inference

Understanding earthquake dynamics is essential for seismic hazard assessment and risk mitigation. In this context, Bayesian inference yields valuable insights into model parameters through a combination of simulation models and real-world data. Such Bayesian parameter inference with uncertainty quantification (UQ) requires numerous simulation runs and is therefore often computationally out of reach. Already a single high-fidelity earthquake simulation – governed by a linear hyperbolic seismic wave equation coupled nonlinearly to a friction law – is computationally expensive. In this study, we investigate the use of fused ensemble simulations as a means to accelerate large earthquake simulation workflows and UQ studies. Via fused ensembles, we turn the element-local small sparse/ dense matrix operations into tensor contractions working on a dense rank-3 tensor and sparse matrices. These are executed with better computational efficiency on CPUs, due to better exploitation of SIMD instructions. We implement and evaluate this approach in SeisSol, a high-performance computing software for simulation of complex earthquake events, which implements an Arbitrary high-order DERivative Discontinuous Galerkin (ADER-DG) scheme. Our results demonstrate that fused simulations can be up to 4.56 times faster than with single execution – though depending strongly on discretization order, size of the problem and compute architecture. For a full UQ example workflow, we demonstrate savings of 35% of node hours for the entire workflow.

This site is open source. Improve this page.