# Self-Supervised Adversarial Example Detection by Disentangled Representation

arXiv: https://arxiv.org/abs/2105.03689 · HTML: https://mlsanity.com/paper/2105.03689/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 34 of its 46 references that are in the map.
- First posted on arXiv 8 May 2021.

## Builds on (most cited first, 30 of 34)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Explaining and Harnessing Adversarial Examples](https://mlsanity.com/paper/1412.6572/) (2014)
- [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)
- [Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithms](https://mlsanity.com/paper/1708.07747/) (2017)
- [Towards Evaluating the Robustness of Neural Networks](https://mlsanity.com/paper/1608.04644/) (2016)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [Character-level Convolutional Networks for Text Classification](https://mlsanity.com/paper/1509.01626/) (2015)
- [A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attacks](https://mlsanity.com/paper/1807.03888/) (2018)
- [DeepFool: a simple and accurate method to fool deep neural networks](https://mlsanity.com/paper/1511.04599/) (2015)
- [Adversarial Examples Are Not Bugs, They Are Features](https://mlsanity.com/paper/1905.02175/) (2019)
- [Adversarial Machine Learning at Scale](https://mlsanity.com/paper/1611.01236/) (2016)
- [InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets](https://mlsanity.com/paper/1606.03657/) (2016)
- [Memorizing Normality to Detect Anomaly: Memory-augmented Deep Autoencoder for Unsupervised Anomaly Detection](https://mlsanity.com/paper/1904.02639/) (2019)
- [Adversarial Examples Are Not Easily Detected: Bypassing Ten Detection Methods](https://mlsanity.com/paper/1705.07263/) (2017)
- [Opening the Black Box of Deep Neural Networks via Information](https://mlsanity.com/paper/1703.00810/) (2017)
- [Generating Natural Language Adversarial Examples](https://mlsanity.com/paper/1804.07998/) (2018)
- [Characterizing Adversarial Subspaces Using Local Intrinsic Dimensionality](https://mlsanity.com/paper/1801.02613/) (2018)
- [MagNet: a Two-Pronged Defense against Adversarial Examples](https://mlsanity.com/paper/1705.09064/) (2017)
- [Generating Adversarial Examples with Adversarial Networks](https://mlsanity.com/paper/1801.02610/) (2018)
- [Detecting Adversarial Samples from Artifacts](https://mlsanity.com/paper/1703.00410/) (2017)
- [On the (Statistical) Detection of Adversarial Examples](https://mlsanity.com/paper/1702.06280/) (2017)
- [Towards the first adversarially robust neural network model on MNIST](https://mlsanity.com/paper/1805.09190/) (2018)
- [SafetyNet: Detecting and Rejecting Adversarial Examples Robustly](https://mlsanity.com/paper/1704.00103/) (2017)
- [The Odds are Odd: A Statistical Test for Detecting Adversarial Examples](https://mlsanity.com/paper/1902.04818/) (2019)
- [MagNet and "Efficient Defenses Against Adversarial Attacks" are Not Robust to Adversarial Examples](https://mlsanity.com/paper/1711.08478/) (2017)
- [Early Methods for Detecting Adversarial Images](https://mlsanity.com/paper/1608.00530/) (2016)
- [Detecting and Diagnosing Adversarial Images with Class-Conditional Capsule Reconstructions](https://mlsanity.com/paper/1907.02957/) (2019)

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

- [Character-Level Perturbations Disrupt LLM Watermarks](https://mlsanity.com/paper/2509.09112/) (2025)
- [Evaluating Membership Inference Through Adversarial Robustness](https://mlsanity.com/paper/2205.06986/) (2022)
- [Masked Language Model Based Textual Adversarial Example Detection](https://mlsanity.com/paper/2304.08767/) (2023)
- [Are You Still the Agent I Authorized? Earned Authority under a Fixed Ceiling for Evolving Agents](https://mlsanity.com/paper/2607.23586/) (2026)
- [Exploring Gradient-Guided Masked Language Model to Detect Textual Adversarial Attacks](https://mlsanity.com/paper/2504.08798/) (2025)
- [When Better Features Mean Greater Risks: The Performance-Privacy Trade-Off in Contrastive Learning](https://mlsanity.com/paper/2506.05743/) (2025)
- [SETTI: A Self-supervised Adversarial Malware Detection Architecture in an IoT Environment](https://mlsanity.com/paper/2204.07772/) (2022)

## Abstract (arXiv)

Deep learning models are known to be vulnerable to adversarial examples that are elaborately designed for malicious purposes and are imperceptible to the human perceptual system. Autoencoder, when trained solely over benign examples, has been widely used for (self-supervised) adversarial detection based on the assumption that adversarial examples yield larger reconstruction errors. However, because lacking adversarial examples in its training and the too strong generalization ability of autoencoder, this assumption does not always hold true in practice. To alleviate this problem, we explore how to detect adversarial examples with disentangled label/semantic features under the autoencoder structure. Specifically, we propose Disentangled Representation-based Reconstruction (DRR). In DRR, we train an autoencoder over both correctly paired label/semantic features and incorrectly paired label/semantic features to reconstruct benign and counterexamples. This mimics the behavior of adversarial examples and can reduce the unnecessary generalization ability of autoencoder. We compare our method with the state-of-the-art self-supervised detection methods under different adversarial attacks and different victim models, and it exhibits better performance in various metrics (area under the ROC curve, true positive rate, and true negative rate) for most attack settings. Though DRR is initially designed for visual tasks only, we demonstrate that it can be easily extended for natural language tasks as well. Notably, different from other autoencoder-based detectors, our method can provide resistance to the adaptive adversary.