[New Model] <LBGAT>

[jiequancui]

Paper Information

• Learnable Boundary Guided Adversarial Training:
• https://arxiv.org/abs/2011.11164:
• Jiequan Cui, Shu Liu, Liwei Wang, Jiaya Jia:

Leaderboard Claim(s)

Add here the claim for your model(s). Copy and paste the following subsection for the number of models you want to add.

Model 1

• Venue: "ICCV2021"
• Architecture: WideResNet34-10
• Dataset: CIFAR-100
• Threat Model: Linf
• eps: 8/255
• Clean accuracy: 62.99
• Robust accuracy: 31.20
• Additional data: false
• Evaluation method: AutoAttack
• Checkpoint and code: model, log and code.

Model Zoo:

• [ Y] I want to add my models to the Model Zoo (check if true)
• [ Y] I use an architecture that is included among those here or in timm. If not, I added the link to the architecture implementation so that it can be added.
• [ Y] I agree to release my model(s) under MIT license (check if true) OR under a custom license, located here: (put the custom license URL here if a custom license is needed. If no URL is specified, we assume that you are fine with MIT)

[fra31]

Hi,

thanks for the submission, we'll add the model as soon as possible!

[fra31]

Is there any specific difference to the models currently in the leaderboard (so that we can add a footnote to distinguish them)? Or is this meant to replace one of those?

[jiequancui]

We integrate AWP into our LBGAT method. Also, The model is trained with a perturbation size of 8/255.

[fra31]

Added the model with https://github.com/RobustBench/robustbench/pull/125, please let me know if there's something to adjust.

[jiequancui]

Thanks！

[jiequancui]

Hi, will the new model be updated on the leaderboard later?

[fra31]

It should be updated now https://github.com/RobustBench/robustbench.github.io/commit/e5f57c9d6a67cc5adcb803065cad312b564da78f.