EAD: Elastic-Net Attacks to Deep Neural Networks via Adversarial Examples

2017/09/13 by Pin‐Yu Chen, Chen, Pin-Yu, Yash Sharma +7 · 8 citations
Computer Science · #Adversarial Robustness in Machine Learning #Cryptography and Security (cs.CR) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML)

paper · pdf · doi:10.48550/arxiv.1709.04114

openalex publication_date 2017/09/13 · openalex created_date 2017/09/25 · openalex updated_date 2026/07/28

Abstract

Recent studies have highlighted the vulnerability of deep neural networks (DNNs) to adversarial examples - a visually indistinguishable adversarial image can easily be crafted to cause a well-trained model to misclassify. Existing methods for crafting adversarial examples are based on L₂ and L_∞ distortion metrics. However, despite the fact that L₁ distortion accounts for the total variation and encourages sparsity in the perturbation, little has been developed for crafting L₁-based adversarial examples. In this paper, we formulate the process of attacking DNNs via adversarial examples as an elastic-net regularized optimization problem. Our elastic-net attacks to DNNs (EAD) feature L₁-oriented adversarial examples and include the state-of-the-art L₂ attack as a special case. Experimental results on MNIST, CIFAR10 and ImageNet show that EAD can yield a distinct set of adversarial examples with small L₁ distortion and attains similar attack performance to the state-of-the-art methods in different attack scenarios. More importantly, EAD leads to improved attack transferability and complements adversarial training for DNNs, suggesting novel insights on leveraging L₁ distortion in adversarial machine learning and security implications of DNNs.

Citations

• Adam: A Method for Stochastic Optimization
• Distilling the Knowledge in a Neural Network
• On the (Statistical) Detection of Adversarial Examples
• Towards Interpretable Deep Neural Networks by Leveraging Adversarial Examples
• Robust Physical-World Attacks on Deep Learning Models
• SafetyNet: Detecting and Rejecting Adversarial Examples Robustly
• Improving the Robustness of Deep Neural Networks via Stability Training
• Universal adversarial perturbations

Cited by

• When Robots Obey the Patch: Universal Transferable Patch Attacks on Vision-Language-Action Models
• An Investigation of Visual Foundation Models Robustness
• Evaluating the Evaluators: Trust in Adversarial Robustness Tests
• Neural Network Reprogrammability: A Unified Theme on Model Reprogramming, Prompt Tuning, and Prompt Instruction
• Conformal Mixed-Integer Constraint Learning with Feasibility Guarantees
• On The Utility of Conditional Generation Based Mutual Information for Characterizing Adversarial Subspaces
• Adversarial Detection and Correction by Matching Prediction Distributions
• GenoArmory: A Unified Evaluation Framework for Adversarial Attacks on Genomic Foundation Models

Related

• Attacking the Madry Defense Model with L₁-based Adversarial Examples
• Generating Unrestricted Adversarial Examples via Three Parameters
• An ADMM-Based Universal Framework for Adversarial Attacks on Deep Neural Networks
• AFLOW: Developing Adversarial Examples under Extremely Noise-limited Settings
• Enhancing Adversarial Transferability with Adversarial Weight Tuning