# Noise Modulation: Let Your Model Interpret Itself

arXiv: https://arxiv.org/abs/2103.10603 · HTML: https://mlsanity.com/paper/2103.10603/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 27 of its 34 references that are in the map.
- First posted on arXiv 19 Mar 2021.

## Builds on (most cited first, 27 of 27)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Towards Deep Learning Models Resistant to Adversarial Attacks](https://mlsanity.com/paper/1706.06083/) (2017)
- [Intriguing properties of neural networks](https://mlsanity.com/paper/1312.6199/) (2013)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples](https://mlsanity.com/paper/1802.00420/) (2018)
- [Sanity Checks for Saliency Maps](https://mlsanity.com/paper/1810.03292/) (2018)
- [SmoothGrad: removing noise by adding noise](https://mlsanity.com/paper/1706.03825/) (2017)
- [Robustness May Be at Odds with Accuracy](https://mlsanity.com/paper/1805.12152/) (2018)
- [Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods](https://mlsanity.com/paper/1705.07263/) (2017)
- [Object Detection in 20 Years: A Survey](https://mlsanity.com/paper/1905.05055/) (2019)
- [Adversarially Robust Generalization Requires More Data](https://mlsanity.com/paper/1804.11285/) (2018)
- [Not Just a Black Box: Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1605.01713/) (2016)
- [Adversarial Examples for Semantic Segmentation and Object Detection](https://mlsanity.com/paper/1703.08603/) (2017)
- [Improving the Adversarial Robustness and Interpretability of Deep Neural Networks by Regularizing their Input Gradients](https://mlsanity.com/paper/1711.09404/) (2017)
- [Feature Purification: How Adversarial Training Performs Robust Deep Learning](https://mlsanity.com/paper/2005.10190/) (2020)
- [On the Connection Between Adversarial Robustness and Saliency Map Interpretability](https://mlsanity.com/paper/1905.04172/) (2019)
- [Interpreting Adversarially Trained Convolutional Neural Networks](https://mlsanity.com/paper/1905.09797/) (2019)
- [Jacobian Adversarially Regularized Networks for Robustness](https://mlsanity.com/paper/1912.10185/) (2019)
- [What it Thinks is Important is Important: Robustness Transfers through Input Gradients](https://mlsanity.com/paper/1912.05699/) (2019)
- [Towards Interpretable Deep Neural Networks by Leveraging Adversarial Examples](https://mlsanity.com/paper/1708.05493/) (2017)
- [Bridging Adversarial Robustness and Gradient Interpretability](https://mlsanity.com/paper/1903.11626/) (2019)
- [A Survey on Deep Learning Methods for Semantic Image Segmentation in Real-Time](https://mlsanity.com/paper/2009.12942/) (2020)

## Abstract (arXiv)

Given the great success of Deep Neural Networks(DNNs) and the black-box nature of it,the interpretability of these models becomes an important issue.The majority of previous research works on the post-hoc interpretation of a trained model.But recently, adversarial training shows that it is possible for a model to have an interpretable input-gradient through training.However,adversarial training lacks efficiency for interpretability.To resolve this problem, we construct an approximation of the adversarial perturbations and discover a connection between adversarial training and amplitude modulation. Based on a digital analogy,we propose noise modulation as an efficient and model-agnostic alternative to train a model that interprets itself with input-gradients.Experiment results show that noise modulation can effectively increase the interpretability of input-gradients model-agnosticly.