A^4D detects adversarial attacks in an attack- and classifier-agnostic way by measuring non-arbitrary shifts in CLIP embedding space from prompt-based similarity scores.
Torchattacks: A pytorch repository for adversarial attacks
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
abstract
Torchattacks is a PyTorch library that contains adversarial attacks to generate adversarial examples and to verify the robustness of deep learning models. The code can be found at https://github.com/Harry24k/adversarial-attacks-pytorch.
citation-role summary
method 1
citation-polarity summary
roles
method 1polarities
use method 1representative citing papers
Adversarial perturbations possess an inherently low-rank structure that enables more efficient and effective black-box adversarial attacks via subspace projection.
CURE is the first multi-norm certified training method that improves union robustness across l_p norms and unseen perturbations on MNIST, CIFAR-10 and TinyImagenet.
Learning a backdoored reference model as a poisonous-sample oracle enables near-perfect training-time backdoor removal with negligible natural-accuracy loss.
TAME uses a Mixture-of-Experts prompt bank with input-dependent routing and three unsupervised objectives to adaptively defend CLIP against adversarial attacks at inference time, achieving at least 49.1% robustness gain on 11 datasets.
A preprocessor of Gaussian noise plus bilateral filtering yields supralinear adversarial robustness in CNNs and, when paired with adversarial training, ranks near the top of RobustBench while using far less compute, parameters, epochs, and data than prior defenses.
A CNN-plus-quantum-circuit classifier with learned fusion reports lower attack success rates and much higher attack-generation cost than a CNN baseline on MNIST, OrganAMNIST, and CIFAR-10.
Zubov-Net aligns prescribed regions of attraction defined by learnable Lyapunov functions with true regions in Neural ODEs via a differentiable Zubov consistency loss, claiming to reconcile accuracy and certified robustness.
LLM safety evaluations are hindered by noise in dataset curation, automated red-teaming, response generation, and LLM-judge evaluation, making fair comparisons difficult and slowing progress.
citing papers explorer
-
Low Rank Adaptation for Adversarial Perturbation
Adversarial perturbations possess an inherently low-rank structure that enables more efficient and effective black-box adversarial attacks via subspace projection.