The results show that: IMA significantly improves the adversarial robustness of DNNs and outperforms the other defense methods on the evaluated datasets and ...
May 19, 2020 · We propose a novel adversarial training method, named Increasing-Margin Adversarial (IMA) Training, which is supported by an equilibrium state analysis.
Oct 13, 2020 · During training, the IMA method increases the margins of training samples by moving the decision boundaries of the DNN model far away from the training samples ...
We propose a novel adversarial training method, named Increasing-Margin Adversarial (IMA) Training, which is supported by an equilibrium state analysis.
Oct 1, 2023 · Conclusions: Our study has demonstrated that our method can lift the trade-off between standard accuracy and adversarial robustness for the ...
In this study, we propose a novel training method, named In- creasing Margin Adversarial (IMA) Training, to improve DNN robustness against adversarial noises.
Feb 10, 2023 · Our method aims to preserve accuracy while improving robustness by generat- ing optimal adversarial training samples. We evaluate our method and ...
This work investigated the existing adversarial training methods and discovered the challenges that make those methods unsuitable for adaptation in ...
The official source code for the paper "Increasing-Margin Adversarial (IMA) Training to Improve Adversarial Robustness of Neural Networks" ...
People also ask
Does enhanced shape bias improve neural network robustness to common corruptions?
Can we have it all on the trade off between spatial and adversarial robustness of neural networks?
Increasing-Margin Adversarial (IMA) Training to Improve Adversarial Robustness of Neural Networks. L Ma, L Liang. Computer Methods and Programs in Biomedicine ...