⚛️ quantum physics

Noise robust self testing from genuine local operation shared randomness multipartite nonlocality Tests

This paper presents a device-independent, noise-robust self-testing protocol for $N$ -partite genuine nonlocality under local operations and shared randomness, deriving explicit $O(\sqrt{\epsilon})$ bounds for the state and measurements based on the Bell-score deficit.

Original authors: Som Kanjilal

Published 2026-09-15

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Som Kanjilal

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the quantum world, particles can become linked in a way that defies our everyday understanding of how things work. When two or more particles share this connection, known as entanglement, measuring one instantly reveals information about the others, no matter how far apart they are. Scientists have long used this phenomenon to test the very foundations of reality, checking whether the universe follows the strange rules of quantum mechanics or if there is a hidden, simpler explanation we have missed. A central challenge in this field is figuring out how to trust the machines we use to create these links. If a device claims to produce a specific type of entangled state, how can we be sure it is actually doing so without opening the machine up to inspect its internal gears? This is where a concept called "self-testing" comes in. It allows researchers to verify the inner workings of a quantum device solely by observing the patterns of inputs and outputs it produces, treating the machine like a black box that reveals its secrets through its behavior alone.

For years, scientists have been able to self-test simple pairs of particles, but the task becomes exponentially harder as more particles are added to the mix. When dealing with three or more particles, the complexity of possible connections increases, and new types of "hidden" explanations can emerge. Specifically, researchers have struggled to distinguish between a truly unified group of particles acting as a single, inseparable unit and a scenario where smaller groups of particles are linked together in a chain, with the final result just looking like a big group from the outside. A recent study by Som Kanjilal addresses this exact problem. The paper demonstrates a new, highly reliable method to prove that a group of particles is genuinely connected as a whole, rather than being a collection of smaller, independent clusters. This is achieved by using a specific mathematical test that rules out any scenario where the particles could be explained by local interactions between smaller groups, even if those groups share a common source of randomness.

The core of this work involves a specific test designed for a group of N particles, where N is any number greater than or equal to three. The researchers focused on a particular arrangement of particles known as a GHZ state, which is a famous type of entangled state where all particles are perfectly synchronized. In an ideal world, if you perform the right measurements on these particles, you get a perfect score on a specific mathematical inequality. However, real-world experiments are never perfect; they are always subject to noise and imperfections. The challenge has been to create a test that not only identifies the perfect case but also works when the results are slightly off. The author developed a new protocol that acts as a robust self-test. This means that even if the experimental score is slightly lower than the theoretical maximum, the test can still guarantee that the physical state inside the machine is very close to the ideal GHZ state. The method provides a precise mathematical guarantee: if the score is close enough to the top, the actual state of the particles must be within a specific, calculable distance of the target state.

What makes this discovery particularly significant is the strictness of the rules it applies. The test is designed to rule out a broad class of alternative explanations that previous tests might have missed. It specifically excludes models where the particles are not truly all connected at once, but are instead formed by combining smaller groups of particles that share resources locally. Even if these smaller groups are supplemented with a shared source of randomness that all parties can access, the test proves that such a setup cannot reproduce the observed results. This confirms that the correlations seen in the experiment require an irreducible connection involving all N particles simultaneously. The study establishes that the only way to generate these specific patterns of data is through a genuine, multipartite quantum resource that cannot be broken down into simpler, smaller parts.

The researchers achieved this by constructing a detailed mathematical framework that links the observed data directly to the physical state and the measurements being performed. They showed that if the experimental data matches the expected pattern closely enough, there must exist a specific transformation that maps the unknown physical setup to the known ideal GHZ state. This transformation acts like a universal translator, proving that the unknown machine is essentially doing the same thing as the ideal reference, just perhaps with some extra, unused parts attached. The paper provides explicit formulas that quantify exactly how close the physical state is to the ideal one based on how far the experimental score falls short of the maximum. This allows scientists to say with confidence, "If your score is this high, your machine is this close to perfect," without needing to know the internal dimensions or specific design of the device.

Furthermore, the study extends this verification to the measurements themselves. It is not enough to know the particles are in the right state; one must also verify that the devices measuring them are performing the correct actions. The new protocol confirms that the measurement devices are acting exactly as they should, up to a small, calculable error margin that grows only with the square root of the experimental noise. This is a crucial detail because it means the method remains effective even as the noise increases, offering a practical tool for real-world quantum experiments. The results apply to any number of particles greater than or equal to three, making it a scalable solution for future quantum technologies that rely on large networks of entangled particles.

The significance of this work lies in its ability to provide a double layer of certification. On one level, it confirms the quantum nature of the system by proving it cannot be explained by classical physics or simpler quantum models. On another level, it certifies the specific quantum state and measurements being used, ensuring that the hardware is functioning as intended. This dual certification is achieved through a single experiment, making it an efficient and powerful tool for the field. The study does not rely on assumptions about the size of the quantum systems or the specific nature of the measurement devices, making it a truly "device-independent" verification. By ruling out complex causal models involving local compositions of resources, the research solidifies our understanding of what it means for a group of particles to be genuinely connected. It offers a clear, rigorous path forward for validating the complex quantum systems that will underpin future technologies, ensuring that when we claim to have built a network of entangled particles, we truly have.

Technical Summary: Noise Robust Self Testing from Genuine Local Operation Shared Randomness

Problem Statement
Device-independent self-testing allows for the certification of unknown quantum states and measurements solely based on observed input-output statistics, without assumptions regarding the internal dimension of the Hilbert space or the characterization of measurement devices. While robust self-testing protocols exist for standard Bell locality and Svetlichny-type genuine multipartite nonlocality, there is a gap in certification methods for genuine Local Operation and Shared Randomness (LOSR) multipartite nonlocality.

LOSR multipartite nonlocality is a theory-agnostic notion where correlations cannot be generated by local operations acting on arbitrary causal resources shared by at most $N-1$ parties, even if all $N$ parties have access to global shared randomness. Mao et al. [18] previously introduced an $N$ -party Bell-type inequality capable of witnessing this form of nonlocality. However, while violation of this inequality rules out specific causal-network models, it did not previously establish a self-testing protocol. Specifically, it was unknown whether the quantum state and measurements achieving the maximal violation were unique (up to local isometries) or how to quantify the certification when experimental noise results in sub-maximal violations.

Methodology
The paper addresses this gap by constructing a device-independent self-test for the Mao functional (the $N$ -partite inequality from Ref. [18]). The methodology proceeds in two stages: exact self-testing and noise-robust analytic bounds.

Exact Self-Testing via SOS Decomposition:
The author first establishes that any physical realization achieving the maximal quantum value $\beta^Q_N = 2\sqrt{2} + 2(N-2)$ must satisfy specific operator identities. They employ a Sum-of-Squares (SOS) decomposition of the nonlocality operator $S_N$ . This decomposition reveals that at the maximum, the physical state $|\psi\rangle$ must be annihilated by specific residual operators involving the difference between physical measurements and reference stabilizers.
- The proof utilizes a coordinated polar decomposition to construct unitary counterparts ( $M_{0,0}, M_{0,1}$ ) for the first party's observables ( $A_+$ and $A_-$ ), ensuring they satisfy the necessary anticommutation relations on the state.
- A local SWAP isometry is constructed using a circuit of Hadamard gates and controlled-unitary operations. This isometry maps the physical state and measurements to a reference $N$ -qubit GHZ state and its corresponding stabilizer measurements, proving that the maximal realization is unique up to local isometries and auxiliary degrees of freedom.
Noise-Robust Analytic Bounds:
To address experimental imperfections, the paper derives analytic bounds for the case where the observed Bell score is $\beta^Q_N - \epsilon$ (where $\epsilon$ is the deficit).
- The SOS identity is used to decompose the score deficit into squared norms of residual operators ( $r_+, r_-, e_k$ ).
- These residuals are bounded to quantify the deviation of the extracted state from the target GHZ state and the deviation of physical measurements from the reference Pauli operators.
- The author derives explicit $O(\sqrt{\epsilon})$ vector-norm bounds for the state fidelity and measurement actions, avoiding numerical optimization in favor of closed-form analytic expressions.

Key Contributions and Results

Exact Self-Testing of LOSR Nonlocality: The paper proves that the maximal violation of the Mao inequality uniquely self-tests the $N$ -qubit GHZ state and the specific set of local measurements (stabilizers) that generate it. This is the first self-testing result specifically for the genuine LOSR multipartite nonlocality framework.
Analytic Robustness Bounds: The author derives a noise-robust self-test with a fidelity bound independent of $N$ (for the specific normalization of the functional). Specifically, for a Bell-score deficit $\epsilon$ , the fidelity $F_N$ of the extracted state satisfies:
$F_N \geq \max\{0, 1 - \kappa \epsilon\}$
where $\kappa = \frac{3 + \sqrt{5}}{\sqrt{2}}$ .
Measurement Certification: The paper establishes explicit bounds on the distance between the physical measurement actions and the reference Pauli operators ( $\sigma_z, \sigma_x$ ). For the first party, the error scales as $O(\sqrt{\epsilon})$ , and for subsequent parties, the bounds depend on the specific measurement setting, also scaling with $\sqrt{\epsilon}$ .
Dual Certification: The protocol provides a dual certification. At the theory-agnostic level, a violation above the LOSR bound rules out any causal model composed of resources shared by at most $N-1$ parties. Within quantum theory, a score close to the quantum maximum certifies the physical realization is close to the GHZ reference.

Significance and Claims
The paper claims that the Mao functional serves as a noise-robust, device-independent certification protocol for the $N$ -qubit GHZ state and its stabilizer measurements without assumptions on Hilbert space dimensions.

The author distinguishes their work from existing protocols in two ways:

Operational Meaning: Unlike standard Bell or Svetlichny tests, this protocol certifies nonlocality against a specific class of causal explanations (LOSR models) that allow for overlapping resources and global shared randomness, which are excluded by standard Svetlichny models.
No Network Assumptions: Unlike network-assisted self-testing, this protocol does not assume a specific network topology or the independence of multiple sources. Instead, the exclusion of causal models arises purely from the violation of the inequality itself.

The work concludes that the Mao functional offers a unique "dual certification": it simultaneously rules out lower-order causal simulations (in a theory-agnostic sense) and certifies the proximity to the GHZ state (within quantum theory), all derived from a single experimentally accessible nonlocality functional. The author explicitly notes that the coefficient $\kappa$ is independent of $N$ for the absolute deficit, though the tolerance to specific noise models (like white noise mixing) may still exhibit $N$ -dependence due to the scaling of the quantum maximum.

Technical Summary: Noise Robust Self Testing from Genuine Local Operation Shared Randomness

More like this