SQUASH: A SWAP-Based Quantum Attack to Sabotage Hybrid Quantum Neural Networks

TL;DR

This paper introduces SQUASH, a stealthy quantum attack that sabotages hybrid quantum neural networks by inserting SWAP gates, significantly degrading their classification accuracy without needing access to training data.

Contribution

The paper presents a novel circuit-level attack method, SQUASH, that manipulates the quantum circuit structure to disrupt HQNN performance, highlighting a new vulnerability in quantum machine learning.

Findings

Untargeted SWAP attacks reduce accuracy by up to 74.08%.

Targeted SWAP attacks reduce target class accuracy by up to 79.78%.

SQUASH is highly stealthy and does not require data access.

Abstract

We propose a circuit-level attack, SQUASH, a SWAP-Based Quantum Attack to sabotage Hybrid Quantum Neural Networks (HQNNs) for classification tasks. SQUASH is executed by inserting SWAP gate(s) into the variational quantum circuit of the victim HQNN. Unlike conventional noise-based or adversarial input attacks, SQUASH directly manipulates the circuit structure, leading to qubit misalignment and disrupting quantum state evolution. This attack is highly stealthy, as it does not require access to training data or introduce detectable perturbations in input states. Our results demonstrate that SQUASH significantly degrades classification performance, with untargeted SWAP attacks reducing accuracy by up to 74.08\% and targeted SWAP attacks reducing target class accuracy by up to 79.78\%. These findings reveal a critical vulnerability in HQNN implementations, underscoring the need for more resilient architectures against circuit-level adversarial interventions.