Adversarial-Dropout

Implementation of "Adversarial Dropout for Supervised and Semi-Supervised Learning" by Sungrae Park, Jun-Keon Park, Su-Jin Shin and Il-Chul Moon https://arxiv.org/abs/1707.03631.

Most of the code is based on https://github.com/takerum/vat_tf. I simply added an implementation of Virtual Adversarial Dropout loss to it.

Haven't been able yet to replicate the results published in the paper, I believe my calculation of the Jacobian still has some error, but can't figure out how to do it, please let me know if you have an idea.

Usage

(Copied from https://github.com/takerum/vat_tf)

Preparation of dataset for semi-supervised learning

On CIFAR-10

python cifar10.py --data_dir=./dataset/cifar10/

Semi-supervised Learning without augmentation

On CIFAR-10

python train_semisup.py --dataset=cifar10 --data_dir=./dataset/cifar10/ --log_dir=./log/cifar10/ --num_epochs=500 --epoch_decay_start=460 --method=vad