Uncertainty in Deep Learning

Some code (TensorFlow) based on the paper:

A Kendall, Y Gal, “What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?”, NIPS 2017 arXiv

DISCLAIMER: This is NOT the official repo. It is just based on my understanding of the paper. Any feedback is welcome.

I am training a simple autoencoder (regression) to reconstruct MNIST digits.

Getting MNIST

Download MNIST:

./download.sh

Rescale and save in a python dictionary (possibly resize):

python prepro.py

License

This repository is released under the MIT LICENSE.