MSEE: Short Course on Uncertainty Quantification in Physics-based Modeling using Python
The course will be held on May 25-27, 2022. Each day will be broken into three sessions that will run from 9AM-12PM and 1-3PM EDT. During the morning session, instruction will be provided for setting up the day’s activities. In the afternoon, participants will work independently on UQ activities, and the instructors will be available to assist.
This in-person three-day short course will introduce participants to the practice of uncertainty quantification (UQ) in computational modeling of physical systems using the Python programming language. Participants will be introduced to a variety of UQ tasks including uncertainty propagation, surrogate modeling, and Bayesian inference using the UQpy toolbox. By the end of the course, it is the goal that students will have the knowledge to begin applying UQ to computational models in their respective fields of study.
In particular, attendees will learn how to:
- Link a model to the UQpy software to execute UQ tasks.
- Propagate uncertainties through a computational model using simulation-based methods in UQpy
- Construct a surrogate model using polynomial chaos expansions and Gaussian process regression
- Infer the parameters of a model using Bayesian inference
- Participants will be provided with links to video lectures on UQ topics, which they should watch prior to attending the workshop.
- Participants are expected to be proficient in the Python programming language with knowledge that includes Python syntax, data types and structures (e.g. lists and dictionaries), importing and using widely available libraries (e.g. numpy and scipy), and object-oriented programming in Python. Participants who are not already proficient may attend the short course after completing one (or more) of the online courses listed below.
- Participants must bring their own laptop computer with all necessary software installed, including Python 3. Although we will review installation of UQpy and setting the user environment, we will not review Python installation.
- Participants are encouraged to bring an example code that they can use for UQ in their own field of study.