# Enhancing Intrinsic Adversarial Robustness via Feature Pyramid Decoder

arXiv: https://arxiv.org/abs/2005.02552 · HTML: https://mlsanity.com/paper/2005.02552/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 May 2020.

## Built on by (most cited first, 10 of 10)

- [Advances in adversarial attacks and defenses in computer vision: A survey](https://mlsanity.com/paper/2108.00401/) (2021)
- [Adversarial Visual Robustness by Causal Intervention](https://mlsanity.com/paper/2106.09534/) (2021)
- [Robust Real-World Image Super-Resolution against Adversarial Attacks](https://mlsanity.com/paper/2208.00428/) (2022)
- [AdvFilter: Predictive Perturbation-aware Filtering against Adversarial Attack via Multi-domain Learning](https://mlsanity.com/paper/2107.06501/) (2021)
- [Robust Prototypical Few-Shot Organ Segmentation with Regularized Neural-ODEs](https://mlsanity.com/paper/2208.12428/) (2022)
- [General Adversarial Defense Against Black-box Attacks via Pixel Level and Feature Level Distribution Alignments](https://mlsanity.com/paper/2212.05387/) (2022)
- [AFLOW: Developing Adversarial Examples under Extremely Noise-limited Settings](https://mlsanity.com/paper/2310.09795/) (2023)
- [Mitigating Low-Frequency Bias: Feature Recalibration and Frequency Attention Regularization for Adversarial Robustness](https://mlsanity.com/paper/2407.04016/) (2024)
- [Sparta: Spatially Attentive and Adversarially Robust Activation](https://mlsanity.com/paper/2105.08269/) (2021)
- [Identifying Layers Susceptible to Adversarial Attacks](https://mlsanity.com/paper/2107.04827/) (2021)

## Abstract (arXiv)

Whereas adversarial training is employed as the main defence strategy against specific adversarial samples, it has limited generalization capability and incurs excessive time complexity. In this paper, we propose an attack-agnostic defence framework to enhance the intrinsic robustness of neural networks, without jeopardizing the ability of generalizing clean samples. Our Feature Pyramid Decoder (FPD) framework applies to all block-based convolutional neural networks (CNNs). It implants denoising and image restoration modules into a targeted CNN, and it also constraints the Lipschitz constant of the classification layer. Moreover, we propose a two-phase strategy to train the FPD-enhanced CNN, utilizing ε-neighbourhood noisy images with multi-task and self-supervised learning. Evaluated against a variety of white-box and black-box attacks, we demonstrate that FPD-enhanced CNNs gain sufficient robustness against general adversarial samples on MNIST, SVHN and CALTECH. In addition, if we further conduct adversarial training, the FPD-enhanced CNNs perform better than their non-enhanced versions.