Skip to content
Analysis of Adversarial Logit Pairing
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.
datasets Add analysis Jul 26, 2018
.gitignore
LICENSE Update LICENSE Mar 4, 2016
README.md Update README.md Aug 13, 2018
examples.png
model_lib.py
robustml_attack.py Add analysis Jul 26, 2018
robustml_eval.py Add analysis Jul 26, 2018
robustml_model.py

README.md

Evaluating and Understanding the Robustness of Adversarial Logit Pairing

The code in this repository, forked from the official implementation, evaluates the robustness of Adversarial Logit Pairing, a proposed defense against adversarial examples.

On the ImageNet 64x64 dataset, with an L-infinity perturbation of 16/255 (the threat model considered in the original paper), we can make the classifier accuracy 0.1% and generate targeted adversarial examples (with randomly chosen target labels) with 98.6% success rate using the provided code and models.

See our writeup here for our analysis, including visualizations of the loss landscape induced by Adversarial Logit Pairing.

Pictures

Obligatory pictures of adversarial examples (with randomly chosen target classes).

Setup

Download and untar the ALP-trained ResNet-v2-50 model into the root of the repository.

RobustML evaluation

Run with:

python robustml_eval.py --imagenet-path <path>
You can’t perform that action at this time.