Notebooks and Scripts for Uncertainty Quantification in ML models
Each of the .ipynb files in the notebooks
folder walks you through one application of a UQ method and leverages some pre-written functions from Utility_Scripts
to do so.
There are six primary methods that these notebooks explore (along with some WIPs): 0. Ensemble methods
- Training on Uncertainty
- PyMC3 metthods
- Tensorflow Probability (Regression) (WIP)
- Tensorflow Probability (Classification) (WIP)
- Using Dropout in NNs