Skip to content
Low-variance and unbiased gradient for backpropagation through categorical random variables, with application in variational auto-encoder and reinforcement learning. ICML 2019
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
rl
toy
vae
README.md
README.txt

README.md

Code to show the simulation results in [ARSM: Augment-REINFORCE-Swap-Merge Estimator for Gradient Backpropagation Through Categorical Variables]

Data sets

The MNIST data is self-contained and the Omniglot data is in the repository.
RL data is from OpenAI Gym

Citations

Below are the paper to cite if you find the algorithms in this repository useful in your own research:

@inproceedings{ARSM_ICML2019,
title={{ARSM}: {A}ugment-{REINFORCE}-swap-merge estimator for gradient backpropagation through categorical variables},
author={Mingzhang Yin and Yuguang Yue and Mingyuan Zhou}, booktitle={ICML}, year={2019} }

License Info

This code is offered under the MIT License.

You can’t perform that action at this time.