Experimental implementation of dimension-dependent contextuality\n inequality

Emil Håkansson, Amélie Piveteau, Alban Seguinard, Muhammad Sadiq, Mohamed Bourennane, Otfried Gühne, Martin PlávalaarXiv 2024 · DOI · arXiv

Abstract

We present a derivation and experimental implementation of a dimension-dependent contextuality inequality to certify both the quantumness and dimensionality of a given system. Existing methods for certification of the dimension of quantum system can be cheated by using larger classical systems, creating a potential loophole in these benchmarks. Our approach uses contextuality inequalities that cannot be violated by classical systems thus closing the previous loophole. We validate this framework experimentally with photons, observing violations of a CHSH-based contextuality inequality and surpassing the qutrit bound of the CGLMP4-based contextuality inequality. These show that contextuality can be used for noise-robust tests of the number of qubits.

Introduction

Verifying the dimensionality of a quantum system—specifically, the number of qubits it can access—is a fundamental challenge in quantum information. Traditional "dimension witnesses" often suffer from a classical loophole: a sufficiently large classical system can mimic the statistics of a lower-dimensional quantum system, making it difficult to distinguish true quantum advantage from classical simulation.

This research addresses this challenge by utilizing contextuality. In quantum mechanics, the outcome of a measurement can depend on the context of other measurements performed simultaneously. Because classical systems are inherently non-contextual, they cannot violate contextuality inequalities. By linking the degree of violation to the system's dimension, the authors provide a benchmark that classical systems cannot cheat.

Theoretical Framework and Methods

The study employs the Preparation and Measurement (P&M) contextuality framework. The methodology involves two primary theoretical advancements:

  1. Mapping Bell to Contextuality: Proving that Bell inequalities (used to test entanglement between distant particles) can be converted into contextuality inequalities for a single system.
  2. Dimension-Dependent Bounds: Demonstrating that the mathematical bounds limiting Bell inequalities based on dimension also apply to these contextuality inequalities.

Experimental Implementation

The framework was validated using single photons. For 1-qubit (2-level) systems, information was encoded in photon polarization. For 2-qubit (4-level) systems, a "hyper-entangled" approach encoded information in both polarization and the spatial path of the photon. A Sagnac interferometer was utilized to create and probe these path-based states within a single particle.

Experimental setup for certifying quantum dimensionality using single photons, featuring a heralded source and a Sagnac interferometer for path encoding.

Data collection involved recording coincidence counts from single-photon detectors ($D_1$ through $D_4$). To determine the maximum theoretical violations under experimental constraints, the team employed a numerical "see-saw" optimization algorithm.

Key Results

The researchers successfully demonstrated violations in two distinct regimes:

  • CHSH Test (Qubit): They observed a violation of $2.8021 \pm 0.0098$, significantly exceeding the classical limit of $2$. This confirms the quantum nature of the system.
  • CGLMP₄ Test (Dimension Certification): This test aimed to distinguish between 3-level (qutrit) and 4-level (2-qubit) systems. While the theoretical bound for a qutrit is 0.315, the experimental 4-level setup achieved a violation of $0.3631 \pm 0.01$. By surpassing the qutrit bound, the experiment certified that the system possessed a dimension of at least 4.

Discussion and Conclusion

A significant finding is that these violations can be observed using "separable" preparations (states without internal entanglement), which are easier to prepare and more robust against noise than entangled states. Furthermore, because the inequalities are linear, they are naturally resilient to the white noise typical in quantum laboratories.

While effective for small systems, scaling these numerical certification methods to larger processors (e.g., 100 qubits) remains a computational challenge. Future work will explore broader applications, such as certifying "no-broadcasting" properties essential for quantum key distribution. This research establishes contextuality as a "gold standard" for the noise-robust certification of quantum dimensionality.

Summary updated

Abstract

We present a derivation and experimental implementation of a\ndimension-dependent contextuality inequality to certify both the quantumness\nand dimensionality of a given system. Existing methods for certification of the\ndimension of quantum system can be cheated by using larger classical systems,\ncreating a potential loophole in these benchmarks. Our approach uses\ncontextuality inequalities that cannot be violated by classical systems thus\nclosing the previous loophole. We validate this framework experimentally with\nphotons, observing violations of a CHSH-based contextuality inequality and\nsurpassing the qutrit bound of the CGLMP4-based contextuality inequality. These\nshow that contextuality can be used for noise-robust tests of the number of\nqubits.\n
BibTeX
@misc{hakansson2024experimental,
  author = {Emil Håkansson and Amélie Piveteau and Alban Seguinard and Muhammad Sadiq and Mohamed Bourennane and Otfried Gühne and Martin Plávala},
  title = {{Experimental implementation of dimension-dependent contextuality\textbackslash\{\}n inequality}},
  year = {2024},
  eprint = {2412.09659},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2412.09659},
}

Loading related content...

Figure