Learnable Boundary Guided Adversarial Training

[jiequancui]

Name: {Learnable Boundary Guided Adversarial Training }

Authors: {Jiequan Cui, Shu Liu, Liwei Wang, Jiaya Jia}

Paper: {https://arxiv.org/abs/2011.11164}

Code: code

Venue: {venue if published}

Does the code implement the robust-ml API and include pre-trained models: {yes}

Dataset: {CIFAR-100}

Threat model: {threat model for which claims are made}

Natural accuracy: {60.55}

Claims: {30.20}

[anishathalye]

I don't see the implementation of the robust-ml API. Am I looking in the wrong place?

[jiequancui]

Hi, are there some examples for PyTorch to implement with robust-ml API?

[anishathalye]

You can see a bunch of examples here: https://www.robust-ml.org/defenses/. Here is one that uses torch: https://github.com/bethgelab/AnalysisBySynthesis. The robust-ml API is independent of the ML library you use, so it shouldn't really make much of a difference between tensorflow and torch etc.