Dear PyMC community,
some colleagues and I are developing UM-Bridge, a language-agnostic interface linking Uncertainty Quantification (UQ) packages and numerical simulation software. It enables
- Speeding up UQ projects from prototype to HPC
- Making UQ and model software available to a wide audience
- Reproducible, community-driven UQ benchmarking
We have specific PyMC support: Any UM-Bridge supporting model running on your system (regardless of language etc) can be connected to from PyTensor like
from umbridge.pymc import UmbridgeOp
op = UmbridgeOp("http://localhost:4242", "posterior")
and then used as a perfectly normal PyTensor Op in PyMC. The UM-Bridge model could be anything from a simple test density all the way to a complex simulation code.
In December we are hosting an online workshop with speakers from Google, digiLab, OpenGoSim Ltd, University of Bath, IFP Energies nouvelles, Karlsruher Institut für Technologie (KIT), Bill & Melinda Gates Foundation, and others. This workshop has the following goals:
- Get users familiar with UM-Bridge
- Initiate collaborations between UQ and simulation experts
- Advance the UQ software ecosystem
The workshop will take place online on 11th and 12th of December 2023.
Registration is available until 1st of December. You find registration form and further details at: https://um-bridge.github.io/workshop/
Let me know if you find this interesting, or might like a deeper integration of UM-Bridge and PyMC!