Bridging Quantum Computing and Differential Privacy: A Survey on Quantum Computing Privacy

Yusheng Zhao, Hui Zhong, Xinyue Zhang, Chi Zhang, Miao Pan

Abstract

While quantum computing has strong potential in data-driven fields, the privacy issue of sensitive or valuable information involved in the quantum algorithm should be considered. Differential privacy (DP), which is a fundamental privacy tool widely used in the classical scenario, has been extended to the quantum domain, i.e., quantum differential privacy (QDP). QDP may become one of the most promising approaches toward privacy-preserving quantum computing since it is not only compatible with classical DP mechanisms but also achieves privacy protection by exploiting unavoidable quantum noise in noisy intermediate-scale quantum (NISQ) devices. This paper provides an overview of the various implementations of QDP and their performance in terms of privacy parameters under the DP setting. Specifically, we propose a taxonomy of QDP techniques, categorizing the literature on whether internal or external randomization is used as a source to achieve QDP and how these implementations are applied to each phase of the quantum algorithm. We also discuss challenges and future directions for QDP.

Introduction

As quantum computing (QC) transitions from theoretical physics to practical applications in finance, drug discovery, and machine learning, a critical vulnerability has emerged: privacy. Quantum algorithms often process sensitive classical data, and the information extracted via measurements can inadvertently leak details about the underlying datasets.

Furthermore, the current "Quantum-as-a-Service" model, where users access hardware via the cloud, exposes data to potential attacks from malicious third parties or cloud providers. While Differential Privacy (DP) is the gold standard for privacy in classical computing, it requires a fundamental rethink for the quantum world because quantum information is stored in "states" (qubits) rather than bits, necessitating new definitions for "neighboring data" and "randomization."

Methodology: A Taxonomy of Quantum Differential Privacy

Quantum Differential Privacy (QDP) quantifies privacy by ensuring that the output of a quantum algorithm is nearly indistinguishable whether a specific individual's data is included in the input or not. The authors propose a structured taxonomy of QDP techniques, categorizing them based on the phase of the quantum algorithm where they are applied:

The general framework of QDP, illustrating how randomization is introduced via encoders, quantum channels, and classical mechanisms to ensure indistinguishability for an attacker.
  1. State Preparation: This involves adding noise to classical data before it is encoded into quantum states or using "randomized encodings" that render the initial quantum state itself private.
  2. Quantum Circuit: This phase leverages the inherent, unavoidable noise in Noisy Intermediate-Scale Quantum (NISQ) devices. Specific "quantum channels"—such as depolarizing or amplitude damping channels—act as natural randomization mechanisms that satisfy QDP requirements.
  3. Quantum Measurement: Privacy is achieved by utilizing "shot noise" (statistical variation from finite experiments) or by applying classical DP mechanisms (like Laplace or Gaussian noise) to the measurement outcomes.

Key Implementations and Privacy Mechanisms

The survey synthesizes various DP-preserving quantum algorithms, highlighting how different randomization sources are integrated into the quantum workflow.

Summary of DP-preserving quantum algorithms across different phases, including their references and implementation methods.

The "Noise-as-a-Feature" Paradigm

In classical computing, noise is an artificial addition to achieve privacy. In the NISQ era, noise is an inherent property of the hardware. QDP turns this hurdle into a feature by "quantifying" existing noise to prove a certain level of privacy without necessarily sacrificing additional performance. For some Quantum Machine Learning (QML) models, the impact of "rotation noise" on performance is surprisingly negligible while still providing robust privacy.

Defining "Quantum Neighbors"

Defining "neighboring quantum states" is a central challenge. The paper discusses various metrics, such as Trace Distance (distinguishability of states) and Quantum Wasserstein Distance, which can provide tighter privacy bounds than traditional metrics.

Key Results and Theoretical Breakthroughs

Recent research has achieved several theoretical milestones in QDP:

  • Tighter Privacy Bounds: Using information-theoretic tools like "quantum hockey-stick divergence," researchers have found more optimistic privacy bounds.
  • Logarithmic Scaling: For many noisy circuits, the privacy budget $\epsilon$ is logarithmically related to the distance $d$ between neighboring states, suggesting high efficiency in privacy costs.
  • Verification as an Eigenvalue Problem: Verifying if a noisy circuit satisfies QDP can be converted into finding the maximum and minimum eigenvalues of a specific matrix, making privacy auditing computationally feasible.

Challenges and Future Directions

The authors identify several "missing links" for the field's maturity:

  1. Standardized Benchmarks: There is a dire need for a "MNIST for QDP" to allow for unified performance and privacy metrics.
  2. Software Integration: Quantum simulators currently lack built-in, "one-click" QDP modules similar to classical libraries like TensorFlow Privacy.
  3. Privacy Accounting: Better tools are needed to track how much privacy is "spent" across multiple iterations of variational algorithms like VQE or QAOA.
  4. Empirical Auditing: Most current work is theoretical; more experimental audits are needed to test mathematical bounds against real-world quantum attacks.

Summary updated

Abstract

While quantum computing has strong potential in data-driven fields, the privacy issue of sensitive or valuable information involved in the quantum algorithm should be considered. Differential privacy (DP), which is a fundamental privacy tool widely used in the classical scenario, has been extended to the quantum domain, i.e., quantum differential privacy (QDP). QDP may become one of the most promising approaches toward privacy-preserving quantum computing since it is not only compatible with classical DP mechanisms but also achieves privacy protection by exploiting unavoidable quantum noise in noisy intermediate-scale quantum (NISQ) devices. This paper provides an overview of the various implementations of QDP and their performance in terms of privacy parameters under the DP setting. Specifically, we propose a taxonomy of QDP techniques, categorizing the literature on whether internal or external randomization is used as a source to achieve QDP and how these implementations are applied to each phase of the quantum algorithm. We also discuss challenges and future directions for QDP. By summarizing recent advancements, we hope to provide a comprehensive, up-to-date review for researchers venturing into this field.
BibTeX
@inproceedings{zhao2024bridging,
  author = {Yusheng Zhao and Hui Zhong and Xinyue Zhang and Chi Zhang and Miao Pan},
  title = {{Bridging Quantum Computing and Differential Privacy: A Survey on Quantum Computing Privacy}},
  booktitle = {QCE},
  year = {2024},
  url = {https://openreview.net/forum?id=JdOfjDLyfE},
}
Figure