Probabilistic Verification of Quantum Devices Under Finite Measurement Resolution and Adversarial Disturbances

Abstract

The development of practical quantum devices is transitioning from laboratory-scale demonstrations to engineered systems intended for integration, deployment, and sustained operation. As quantum hardware increases in complexity and scale, establishing reliable verification procedures under realistic operating conditions becomes a central engineering challenge. Real quantum devices operate under finite measurement resolution, estimator tolerances, drift, hardware constraints, and limited data, all of which fundamentally restrict what can be inferred from experimental observations. Consequently, verification strategies based on static thresholds, ideal measurements, or full microscopic reconstruction are often insufficient for deployment-grade systems.

This thesis develops an engineering-oriented probabilistic framework for modelling and verifying quantum devices under finite measurement resolution and adversarial disturbances. Rather than treating verification as a binary decision, the work reframes it as a system-level inference problem governed by uncertainty, tolerances, and acceptance criteria. Across five manuscript-based studies, the thesis identifies fundamental verification vulnerabilities, introduces probabilistic modelling approaches, and develops practical mitigation, scaling, and governance-oriented strategies compatible with real hardware.

First, the thesis shows that finite measurement resolution can create regions of operational indistinguishability in which conventional quantum integrity checks become statistically non-discriminating, even when the underlying theoretical assumptions remain valid. Building on this result, a probabilistic verification framework is introduced to model acceptance outcomes under uncertainty and to quantify confidence levels rather than rely on fixed thresholds alone.

The thesis then develops an operational threat-modelling framework for adversarial disturbances in continuous-variable quantum communication, classifying structured interference into reconnaissance, exploratory, and denial-of-service regimes on the basis of receiver-observable statistics and finite-sample detectability. In response to such disturbances, phase-first modulation strategies are developed to show that static operating points can be inadequate under structured stress and that lightweight, hardware-compatible adaptations can improve resilience.

To address scalability in large quantum systems, the Effective Mode Approximation is introduced as a reduced-order probabilistic verification framework for collective Hamiltonian behaviour, enabling system-level assessment without full mode-resolved reconstruction. Finally, a probabilistic forecasting framework is developed to model time-dependent cryptographic security degradation under evolving classical and quantum threat capabilities, extending verification concepts to strategic risk assessment and transition planning.

Taken together, these contributions establish a unified probabilistic verification perspective grounded in robustness, scalability, and operational realism. The proposed methods align with established principles from control, signal processing, and system identification, and they provide practical tools for assessing trust, performance, and readiness of quantum hardware as it moves toward real-world deployment.