Code accompanying the ICML 2018 paper High-Quality Prediction Intervals for Deep Learning: A Distribution-Free, Ensembled Approach https://arxiv.org/abs/1802.07167.

How can we get uncertainty estimates from deep learning systems?

_{Estimating model uncertainty.}

_{Comparison against MVE.}

A simple fast demo using Keras is included in QD_AsFastAsPoss_notebook.ipynb.

Main paper code in 5 files:

• main.py
• pso.py
• DataGen.py
• DeepNetPI.py
• utils.py
• inputs.txt

Run main.py to reproduce first figure.

We have included hyperparameters used for the boston and concrete datasets in inputs.txt.