Corpus ID: 53031029

Adversarial Learning and Explainability in Structured Datasets.

@article{Chalasani2018AdversarialLA,
  title={Adversarial Learning and Explainability in Structured Datasets.},
  author={P. Chalasani and S. Jha and Aravind Sadagopan and Xi Wu},
  journal={arXiv: Learning},
  year={2018}
}
We theoretically and empirically explore the explainability benefits of adversarial learning in logistic regression models on structured datasets. In particular we focus on improved explainability due to significantly higher $\textit{feature-concentration}$ in adversarially-learned models: Compared to natural training, adversarial training tends to more efficiently shrink the weights of non-predictive and weakly-predictive features, while model performance on natural test data only degrades… Expand
13 Citations
ML-LOO: Detecting Adversarial Examples with Feature Attribution
  • 16
  • PDF
Bridging Adversarial Robustness and Gradient Interpretability
  • 18
  • PDF
Attribution-driven Causal Analysis for Detection of Adversarial Examples
  • 3
  • PDF
Robust Classification using Robust Feature Augmentation
  • 2
  • PDF
On Relating Explanations and Adversarial Examples
  • 21
  • Highly Influenced
  • PDF
Designing Adversarially Resilient Classifiers using Resilient Feature Engineering
  • 4
  • PDF
Detecting Trojaned DNNs Using Counterfactual Attributions
  • 1
  • PDF
...
1
2
...

References

SHOWING 1-10 OF 47 REFERENCES
Explaining and Harnessing Adversarial Examples
  • 7,020
  • PDF
The Limitations of Deep Learning in Adversarial Settings
  • 2,022
  • PDF
Does Interpretability of Neural Networks Imply Adversarial Robustness?
  • 7
  • PDF
Adversarial Machine Learning at Scale
  • 1,364
  • PDF
Bridging Adversarial Robustness and Gradient Interpretability
  • 18
  • PDF
Adversarial Logit Pairing
  • 281
  • PDF
Adversarial Examples: Attacks and Defenses for Deep Learning
  • 636
  • PDF
...
1
2
3
4
5
...