A Comparative Study of Adversarial Robustness in CNN and CNN-ANFIS Architectures

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Shankar, Kaaustaaub, Dogga, Bharadwaj, Cohen, Kelly
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866908818315673600
author Shankar, Kaaustaaub
Dogga, Bharadwaj
Cohen, Kelly
author_facet Shankar, Kaaustaaub
Dogga, Bharadwaj
Cohen, Kelly
contents Convolutional Neural Networks (CNNs) achieve strong image classification performance but lack interpretability and are vulnerable to adversarial attacks. Neuro-fuzzy hybrids such as DCNFIS replace fully connected CNN classifiers with Adaptive Neuro-Fuzzy Inference Systems (ANFIS) to improve interpretability, yet their robustness remains underexplored. This work compares standard CNNs (ConvNet, VGG, ResNet18) with their ANFIS-augmented counterparts on MNIST, Fashion-MNIST, CIFAR-10, and CIFAR-100 under gradient-based (PGD) and gradient-free (Square) attacks. Results show that ANFIS integration does not consistently improve clean accuracy and has architecture-dependent effects on robustness: ResNet18-ANFIS exhibits improved adversarial robustness, while VGG-ANFIS often underperforms its baseline. These findings suggest that neuro-fuzzy augmentation can enhance robustness in specific architectures but is not universally beneficial.
format Preprint
id arxiv_https___arxiv_org_abs_2602_07028
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Comparative Study of Adversarial Robustness in CNN and CNN-ANFIS Architectures
Shankar, Kaaustaaub
Dogga, Bharadwaj
Cohen, Kelly
Computer Vision and Pattern Recognition
Artificial Intelligence
Convolutional Neural Networks (CNNs) achieve strong image classification performance but lack interpretability and are vulnerable to adversarial attacks. Neuro-fuzzy hybrids such as DCNFIS replace fully connected CNN classifiers with Adaptive Neuro-Fuzzy Inference Systems (ANFIS) to improve interpretability, yet their robustness remains underexplored. This work compares standard CNNs (ConvNet, VGG, ResNet18) with their ANFIS-augmented counterparts on MNIST, Fashion-MNIST, CIFAR-10, and CIFAR-100 under gradient-based (PGD) and gradient-free (Square) attacks. Results show that ANFIS integration does not consistently improve clean accuracy and has architecture-dependent effects on robustness: ResNet18-ANFIS exhibits improved adversarial robustness, while VGG-ANFIS often underperforms its baseline. These findings suggest that neuro-fuzzy augmentation can enhance robustness in specific architectures but is not universally beneficial.
title A Comparative Study of Adversarial Robustness in CNN and CNN-ANFIS Architectures
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2602.07028