QCL-IDS: Quantum Continual Learning for Intrusion Detection with Fidelity-Anchored Stability and Generative Replay

TL;DR

This work addresses the challenge of continual intrusion detection under strict compute, memory, and privacy constraints. It introduces QCL-IDS, a quantum-centric framework that combines Q-FISH stability regularization with privacy-preserving quantum generative replay (QGR), implemented on a data re-uploading variational quantum circuit. The key contributions are (i) Q-FISH, which couples a diagonal Quantum Fisher Information proxy with a fidelity-based functional anchor to limit drift on historical samples, and (ii) privacy-preserving QGR that uses frozen, task-conditioned generator snapshots to synthesize bounded rehearsal data. Through experiments on UNSW-NB15 and CICIDS2017, QCL-IDS achieves the best retention-adaptation trade-off, with gradient-space anchors delivering the strongest mean Attack-F1 and minimal forgetting, demonstrating the viability of quantum-native continual learning in privacy-constrained security pipelines.

Abstract

Continual intrusion detection must absorb newly emerging attack stages while retaining legacy detection capability under strict operational constraints, including bounded compute and qubit budgets and privacy rules that preclude long-term storage of raw telemetry. We propose QCL-IDS, a quantum-centric continual-learning framework that co-designs stability and privacy-governed rehearsal for NISQ-era pipelines. Its core component, Q-FISH (Quantum Fisher Anchors), enforces retention using a compact anchor coreset through (i) sensitivity-weighted parameter constraints and (ii) a fidelity-based functional anchoring term that directly limits decision drift on representative historical traffic. To regain plasticity without retaining sensitive flows, QCL-IDS further introduces privacy-preserved quantum generative replay (QGR) via frozen, task-conditioned generator snapshots that synthesize bounded rehearsal samples. Across a three-stage attack stream on UNSW-NB15 and CICIDS2017, QCL-IDS consistently attains the best retention-adaptation trade-off: the gradient-anchor configuration achieves mean Attack-F1 = 0.941 with forgetting = 0.005 on UNSW-NB15 and mean Attack-F1 = 0.944 with forgetting = 0.004 on CICIDS2017, versus 0.800/0.138 and 0.803/0.128 for sequential fine-tuning, respectively.

Figures (6)

• Figure 1: System overview
• Figure 2: Q-FISH concept schematic
• Figure 3: Replay sampling policy diagram
• Figure 4: Data re-uploading VQC circuit
• Figure 5: The curve of Attack-F1 with epoch for each task(UNSW-NB15)