yaringal/DropoutUncertaintyDemos

Stars
115
Rank 305,916 (Top 7 %)
Language
JavaScript
License
MIT License
Created over 9 years ago
Updated about 9 years ago

yaringal/DropoutUncertaintyDemos

yaringal

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

What My Deep Model Doesn't Know...

What My Deep Model Doesn't Know...

This repository contains the code for the demos in the blog post What My Deep Model Doesn't Know.... ConvnetJS is used here as a framework to interactively demonstrate the properties underlying dropout uncertainty. ConvnetJS was originally developed by Karpathy under the MIT license which is used here as well.

multi-task-learning-example

A multi-task learning example for the paper https://arxiv.org/abs/1705.07115

Jupyter Notebook

DropoutUncertaintyExps

Experiments used in "Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning"

BayesianRNN

Code for the paper "A Theoretically Grounded Application of Dropout in Recurrent Neural Networks"

ConcreteDropout

Code for Concrete Dropout as presented in https://arxiv.org/abs/1705.07832

Jupyter Notebook

DropoutUncertaintyCaffeModels

Dropout As A Bayesian Approximation: Code

HeteroscedasticDropoutUncertainty

Demos demonstrating the difference between homoscedastic and heteroscedastic regression with dropout uncertainty.

VSSGP

Variational Sparse Spectrum Gaussian Process toolkit

CLGP

Categorical Latent Gaussian Process

acquisition_example

Train a simple convnet on the MNIST dataset and evaluate the BALD acquisition function

Jupyter Notebook