Entity · dataset

Fashion-MNIST

datasetactiveprovisionalfashion-mnist-67ace726·1 events·first seen 38h ago

Aliases: Fashion-MNIST

Co-occurring entities

Q-DIBA MNIST Input-Aware Dynamic Backdoor Attack Against Quantum Neural Networks

More like this (12)

MNIST Colored MNIST PneumoniaMNIST Split CIFAR-100 Masked Image Modeling Olivetti Faces Dataset ResNet-18 NAMESAKES: Probing Identity Memorization in Text-to-Image Models ResNet-50 CIFAR-10 ImageNet100 ImageNet-100

Recent events (1)

4arXiv · cs.LG·38h ago·source ↗

Q-DIBA: First input-aware dynamic backdoor attack against Quantum Neural Networks

Researchers introduce Q-DIBA, the first input-aware dynamic backdoor attack targeting Quantum Neural Networks (QNNs), addressing limitations of prior fixed-trigger quantum backdoor methods. The approach jointly trains a classical trigger generator and a victim QNN using a three-mode mini-batch strategy and an ensemble density contrastive loss operating on post-ansatz quantum states before measurement. Experiments on MNIST and Fashion-MNIST demonstrate high attack success rates, stealthiness, and resilience against defenses including spectral-signature detection and fine-tuning. The work highlights a novel security threat relevant to near-term quantum machine learning deployments.

AI Safety Research Q-DIBA MNIST Input-Aware Dynamic Backdoor Attack Against Quantum Neural Networks +1 more