arXiv AI

Adversarial Robustness of Kernel Machines: A Comprehensive Analysis

The paper investigates how resilient kernel-based models are to adversarial attacks. It conducts extensive experiments on binary and multiclass classification tasks, using both white-box and black-box evasion techniques, and compares native kernel machines, random feature classifiers, and Recursive Feature Machines. The study also introduces an algorithm for approximate adversarial training of kernel machines and finds that these models degrade more gracefully than neural networks when faced with adversarial perturbations.

arXiv Machine Learning
Jun 26

Over-parameterization and Adversarial Robustness in Neural Networks: An Overview and Empirical Analysis

arXiv:2406. 10090v3 Announce Type: replace Abstract: Thanks to their extensive capacity, over-parameterized neural networks exhibit superior predictive capabilities and generalization.

By Srishti Gupta, Zhang Chen, Luca Demetrio, Fabio Brau, Xiaoyi Feng, Zhaoqiang Xia, Antonio Emanuele Cin\`a, Maura Pintor, Luca Oneto, Ambra Demontis, Battista Biggio, Fabio Roli
OpenAI Blog
Aug 22, 2019

Testing robustness against unforeseen adversaries

We’ve developed a method to assess whether a neural network classifier can reliably defend against adversarial attacks not seen during training. Our method yields a new metric, UAR (Unforeseen Attack Robustness), which evaluates the robustness of a single model against an unanticipated attack, and highlights the need to measure performance across a more diverse range of unforeseen attacks.

arXiv Machine Learning
Oct 28

Attack logics, not outputs: Concept-based Property Attacks for Shortcut Elimination in Neural Networks

The paper introduces concept-based property attacks, which target logical inconsistencies in human‑interpretable concepts (e.g., “red ∧ octagonal → stop_sign”) rather than merely flipping final class labels. It formalizes these properties, proves that attacking them reduces the search space compared to class‑level attacks, and implements them using concept‑based explainable AI techniques. Empirical results across vision and code‑processing models show that retraining with these attacks preserves robustness gains while better eliminating semantic biases.

By Raik Dankworth, Felix M\"achtle, Thomas Eisenbarth, Gesina Schwalbe
arXiv Machine Learning
Sep 22

Reinforcement Learning Inspired Black-box Adversarial Attacks for Computer Vision

The paper introduces RIBA, a reinforcement‑learning inspired black‑box adversarial attack that generates perturbations for neural networks with fewer queries than existing methods. RIBA achieves a 25.4% reduction in median queries on ResNet‑18/Cifar10 and a 22.5% reduction on Vit‑B/16/ImageNet, while matching white‑box attack performance on an adversarially trained model.

By Florian Krone, Elena Hoemann, Sven Hallerbach
arXiv AI
Jul 9

LipSSD: Lipschitz-Constrained Single-Shot Detection for Adversarially Robust Object Detection

arXiv:2607. 06592v1 Announce Type: cross Abstract: Object detectors have many applications in safety-critical systems, but they are known to be sensitive to worst-case perturbations such as adversarial attacks, which limits their applicability in real-world scenarios.

By Vincent L\'eb\'e (IRIT, DTIPG - SNCF, UT3), Yannick Prudent (IRIT, DTIPG - SNCF, UT3), Corentin Friedrich (IRIT, DTIPG - SNCF, UT3), Thomas Massena (IRIT, DTIPG - SNCF, UT3), Ronan Sicre (IRIT), Franck Mamalet
arXiv Machine Learning
Oct 3

Evasion Attacks: How Adversarial Noise Bypasses ML Classifiers

The paper reports a reproducible study of evasion attacks on image and text classifiers. A compact convolutional network on MNIST achieved 98.63% clean accuracy but dropped to 60.20% under FGSM with ε=0.15 and 1.72% with ε=0.30, while PGD reduced accuracy to 32.47% and 0.41%; a bit‑depth‑reduction defense only partially restored performance. In contrast, a DistilBERT model fine‑tuned on the SMS Spam Collection reached 98.75% accuracy and 94.96% F1‑score, yet a sequence of predefined perturbations produced only modest probability shifts and did not flip spam to ham predictions.

By Parker Hummel (Minot State University), Ryne Skabo (Minot State University), Muhammad Abusaqer (Minot State University)