# When Causal Intervention Meets Adversarial Examples and Image Masking for Deep Neural Networks

arXiv: https://arxiv.org/abs/1902.03380 · HTML: https://mlsanity.com/paper/1902.03380/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 9 Feb 2019.

## Built on by (most cited first, 7 of 7)

- [Certified Robustness Against Natural Language Attacks by Causal Intervention](https://mlsanity.com/paper/2205.12331/) (2022)
- [Adversarial Visual Robustness by Causal Intervention](https://mlsanity.com/paper/2106.09534/) (2021)
- [Training a Resilient Q-Network against Observational Interference](https://mlsanity.com/paper/2102.09677/) (2021)
- [LAFFNet: A Lightweight Adaptive Feature Fusion Network for Underwater Image Enhancement](https://mlsanity.com/paper/2105.01299/) (2021)
- [Saliency Guided Adversarial Training for Learning Generalizable Features with Applications to Medical Imaging Classification System](https://mlsanity.com/paper/2209.04326/) (2022)
- [Causal Discovery in Knowledge Graphs by Exploiting Asymmetric Properties of Non-Gaussian Distributions](https://mlsanity.com/paper/2106.01043/) (2021)
- [Neural Networks Decoded: Targeted and Robust Analysis of Neural Network Decisions via Causal Explanations and Reasoning](https://mlsanity.com/paper/2410.05484/) (2024)

## Abstract (arXiv)

Discovering and exploiting the causality in deep neural networks (DNNs) are crucial challenges for understanding and reasoning causal effects (CE) on an explainable visual model. "Intervention" has been widely used for recognizing a causal relation ontologically. In this paper, we propose a causal inference framework for visual reasoning via do-calculus. To study the intervention effects on pixel-level features for causal reasoning, we introduce pixel-wise masking and adversarial perturbation. In our framework, CE is calculated using features in a latent space and perturbed prediction from a DNN-based model. We further provide the first look into the characteristics of discovered CE of adversarially perturbed images generated by gradient-based methods footnote~~https://github.com/jjaacckkyy63/Causal-Intervention-AE-wAdvImg. Experimental results show that CE is a competitive and robust index for understanding DNNs when compared with conventional methods such as class-activation mappings (CAMs) on the Chest X-Ray-14 dataset for human-interpretable feature(s) (e.g., symptom) reasoning. Moreover, CE holds promises for detecting adversarial examples as it possesses distinct characteristics in the presence of adversarial perturbations.