⚛️ quantum physics

Testing bosonic and fermionic Gaussian states with a constant number of copies

This paper demonstrates that a constant number of copies, independent of system size, suffices to test whether a pure bosonic or fermionic state is Gaussian by leveraging state symmetries and spectral gap analysis.

Original author: Jop Briët, Davi Castro-Silva, Michael Walter, Freek Witteveen

Published 2026-11-10

📖 6 min read🧠 Deep dive

CC BY 4.0

Original author: Jop Briët, Davi Castro-Silva, Michael Walter, Freek Witteveen

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 author. For technical accuracy, refer to the original paper. Read full disclaimer

In the vast landscape of quantum physics, where particles behave in ways that defy our everyday intuition, certain states of matter stand out for their predictability and utility. These are known as Gaussian states. They are the workhorses of the quantum world, appearing as the natural resting states of many physical systems, from the light in a laser to the vibrations in a crystal. Because they are mathematically well-behaved, scientists often use them as a foundation for building complex quantum technologies, such as secure communication networks and powerful computers. However, the real world is messy. Quantum systems are easily disturbed, and the delicate states required for advanced computing can easily degrade into something more chaotic and less useful. This creates a critical need for a way to check the health of these systems: given a collection of quantum particles, can we quickly determine if they are still in that useful, orderly Gaussian state, or have they drifted too far away?

For years, scientists faced a frustrating hurdle in answering this question. While they knew how to test for Gaussian states, the method required a number of checks that grew larger and larger as the system itself grew. If a system had just a few particles, a few checks might suffice. But as the system scaled up to the hundreds or thousands of particles needed for real-world applications, the number of required checks became impractical, effectively making the test impossible for large machines. The prevailing assumption was that the complexity of the system would always demand a proportional increase in testing effort. This paper challenges that assumption directly. The researchers demonstrate that it is possible to test whether a quantum state is Gaussian using a fixed, small number of copies, regardless of how large the system is. They have found a way to verify the order of a quantum system with a constant amount of effort, a discovery that removes a major barrier to scaling up quantum technologies.

The team, led by researchers from institutions in the Netherlands, Germany, and the United Kingdom, approached the problem by looking at the unique symmetries that define Gaussian states. In the quantum realm, symmetry is a powerful indicator of order. Gaussian states possess a specific kind of rotational symmetry when viewed through the lens of their mathematical description. The researchers realized that if you take two copies of a quantum state and perform a specific measurement that checks for this symmetry, a true Gaussian state will always pass the test. The challenge, however, was not just to identify perfect states, but to understand what happens when a state is only almost Gaussian. If a state fails the test slightly, how far is it from being perfect? In many scientific contexts, a small failure in a test might hide a massive deviation in the underlying reality, especially as systems get bigger.

To solve this, the authors developed a new mathematical framework that connects the probability of passing a test to the actual distance of the state from the ideal Gaussian form. They treated the problem as a journey across a landscape of possibilities, where the height of the terrain represented how likely a state was to pass the test. By analyzing the slope of this landscape, they could prove that if a state is close enough to passing, it must be physically close to a true Gaussian state. This logic holds true even as the number of particles in the system increases. The key to their success was linking their test to a known mathematical concept involving random walks on a sphere, a problem that had been studied extensively in pure mathematics. By borrowing the known results from that field, they were able to prove that the "slope" of their test landscape is steep enough to guarantee that a high pass rate means the state is genuinely close to the target, with no dependence on the system's size.

The paper details specific tests for different types of quantum systems, including those made of fermions, which are particles like electrons, and bosons, which include particles of light. For fermionic systems, the test involves taking two copies of the state and checking if they remain unchanged when rotated relative to each other in a specific way. For bosonic systems, the approach is similar but requires slightly different handling of the particles' positions and momenta. In all cases, the researchers showed that the number of copies needed to achieve a desired level of confidence depends only on how precise the user wants to be, not on the size of the machine. For instance, to be highly confident that a state is not Gaussian, one might need to repeat the test a specific number of times, but that number remains constant whether the system has ten modes or ten thousand.

This finding is significant because it transforms a theoretical possibility into a practical tool. Previously, the cost of verifying a quantum state was seen as a bottleneck that would grow with the complexity of the technology. Now, the authors have shown that this cost is fixed. They provide explicit instructions for how to perform these tests using standard quantum operations, such as beam splitters, which are devices that mix light beams. The tests are designed to be robust; they do not require perfect equipment or flawless conditions to work, only that the state is sufficiently far from being Gaussian to be detected. The researchers also extended their methods to test for coherent states, a specific type of Gaussian state often used in lasers, and for states with a specific average value, showing the versatility of their approach.

The work stands as a rigorous proof rather than a simulation or a suggestion. The authors have mathematically demonstrated that their tests work with a constant number of copies, providing a guarantee that holds for any system size. They did not merely suggest that this might be possible; they proved it by combining the physical properties of quantum states with deep results from probability theory. This result resolves a long-standing open question in the field of quantum information, confirming that the sample complexity of testing Gaussianity does not scale with the system size. It offers a clear path forward for engineers and scientists building large-scale quantum devices, assuring them that they can verify the quality of their systems without being overwhelmed by the sheer number of particles involved.

In the end, this research provides a fundamental reassurance for the future of quantum technology. It shows that the tools needed to monitor and maintain the delicate order of quantum states do not need to become impossibly complex as the technology grows. By leveraging the inherent symmetries of nature and applying sophisticated mathematical insights, the researchers have carved out a space where efficiency and reliability can coexist, even in the most complex quantum systems. Their work ensures that as we build larger and more powerful quantum machines, we will not lose the ability to know what they are doing, keeping the door open for the next generation of quantum discoveries.

Technical Summary: Testing Bosonic and Fermionic Gaussian States with a Constant Number of Copies

Problem Statement
Bosonic and fermionic Gaussian quantum states are fundamental in quantum physics and information processing, serving as ground states for quadratic Hamiltonians and resources for classical simulation. Conversely, non-Gaussian states are essential for quantum advantage beyond classical simulability. A central question in quantum property testing is whether one can determine if a given pure state $|\psi\rangle$ is Gaussian or is bounded away from the set of pure Gaussian states using a measurement on a finite number of copies. Previous work provided exact characterizations but established sample complexity bounds that scaled with the number of modes $m$ (system size), leaving open whether a constant number of copies suffices independent of $m$ .

Methodology
The authors develop a testing framework based on the symmetry properties of Gaussian states under phase-space rotations. The core methodology involves three main components:

Symmetry-Based Tests:
- Fermionic and Mean-Zero Bosonic States: The tests utilize a two-copy observable. For fermions, the operator $\Lambda = 2 \sum (a_j^\dagger \otimes a_j + a_j \otimes a_j^\dagger)$ generates rotations between copies. For mean-zero bosons, $\Lambda = i \sum (a_j \otimes a_j^\dagger - a_j^\dagger \otimes a_j)$ serves a similar role. The test accepts if the state is in the kernel of $\Lambda$ , which corresponds to the subspace invariant under phase-space rotations (beam splitters of arbitrary angles).
- General Bosonic States: To handle non-zero means, the authors apply a balanced beam splitter to two copies and retain the difference output, which has a vanishing mean. This reduces the general case to the mean-zero test.
- Coherent States: A specific two-copy test involving a beam splitter and a vacuum check is employed.
- General Bosonic Gaussian States: A four-copy test is constructed by applying the mean-zero test to the difference outputs of two balanced beam splitters.
Robustness Analysis via Gradient Flows:
To prove that high acceptance probability implies proximity to the set of Gaussian states (robustness), the authors analyze the gradient flow of the acceptance probability function $p(\psi)$ . They establish a lower bound on the gradient norm proportional to the rejection probability ( $1-p$ ). This approach relies on a quantitative Łojasiewicz inequality, linking the geometry of the acceptance probability landscape to the distance from the target set.
Spectral Gap Connection:
The gradient bound is derived by analyzing the spectrum of a three-copy operator $S = \frac{1}{3}(P_{12} + P_{13} + P_{23})$ , where $P_{ab}$ projects onto the rotation-invariant subspace of copies $a$ and $b$ . The authors demonstrate that $S$ is unitarily equivalent to the transition operator of the Kac random walk on the sphere $S^2$ . By leveraging known spectral gap results for the Kac walk (specifically a gap of $5/12$ ), they bound the spectral gap of $S$ independently of the number of modes $m$ .

Key Contributions and Results
The primary contribution is the proof that testing Gaussianity requires a constant number of copies, independent of the system size $m$ .

Theorem 1.1 (Main Result): Both bosonic and fermionic systems admit tests using $O(\frac{1}{\varepsilon^2} \log \frac{1}{\delta})$ $O (\frac{1}{ε ^{2}} lo g \frac{1}{δ})$ copies to achieve the following:
- If $|\psi\rangle$ is Gaussian, the test accepts with probability 1 (perfect completeness).
- If $|\psi\rangle$ is at trace distance $\varepsilon$ from all Gaussian states (i.e., $F_G^2(\psi) \leq 1-\varepsilon^2$ ), the test rejects with probability at least $1-\delta$ .
Specific Tests:
- Fermionic Gaussianity: Uses 2 copies.
- Mean-zero Bosonic Gaussianity: Uses 2 copies.
- Coherent States: Uses 2 copies.
- General Bosonic Gaussianity: Uses 4 copies.
Theoretical Framework: The paper introduces a general method (Theorem 1.2/3.4) for establishing robustness of property tests with perfect completeness by relating the spectral gap of a symmetrized multi-copy operator to the gradient of the acceptance probability. This connects quantum property testing to the spectral theory of random walks (Kac walk).

Significance and Claims
The paper claims to resolve the open problem of the sample complexity for Gaussianity testing, demonstrating that it is constant with respect to the number of modes. This contrasts with previous bounds that scaled as $\tilde{O}(m \varepsilon^{-2})$ or required learning the state.

The authors emphasize that their tests exploit the enhanced symmetry of Gaussian states (invariance under orthogonal transformations in the multi-copy space, related to Howe–Kashiwara–Vergne duality) rather than estimating covariance matrices. The robustness analysis is global, extending previous results that were limited to states close to Gaussian. The work also highlights a novel connection between quantum testing protocols and the spectral gap of the Kac random walk, utilizing a gradient flow argument to translate spectral properties into geometric bounds on state fidelity.

The paper notes that the results apply to pure states and that the tests have perfect completeness. The authors also acknowledge that the core insight regarding the spectral gap bound was facilitated by an LLM, while the formulation, conjecture, and exposition were human contributions.

More like this