🔬 physics

Locality Spectroscopy of Quantum Statistical Information

This paper introduces locality spectroscopy, a coordinate-invariant framework that links subsystem information structures to Fisher information scaling under local noise, enabling the characterization of how quantum parameter information becomes locally identifiable in composite systems through reanalysis of logical-qubit experiments.

Original authors: Idrees Oreibi, Rehab Shather Abdul Hamza

Published 2026-09-17

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Idrees Oreibi, Rehab Shather Abdul Hamza

Original paper licensed under CC BY 4.0 (https://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, information is not a single, solid object sitting in one place. Instead, it is a delicate pattern woven across many particles, where the whole can know something that no single part can reveal on its own. Scientists have long known how to measure the total amount of information a system holds, but a deeper question has remained: exactly where does that information first become visible? If you were to look at a small group of particles within a larger machine, at what size does that group become large enough to show you the secret the whole system is keeping? Answering this is crucial for building better quantum sensors and computers, because it tells engineers how much of a system they need to control to get a useful reading, and how much noise they can tolerate before the signal disappears.

A team of researchers has now mapped out this hidden landscape, creating a new way to see how information spreads through a quantum system. They call this process "locality spectroscopy." Imagine trying to understand a complex story by reading only small snippets of text. If the story is written in a way that the meaning only appears when you read three words together, looking at one or two words will tell you nothing. The researchers developed a method to find that specific "three-word" threshold for any quantum system. They defined a spectrum that lists the smallest size of a particle group needed to identify each independent piece of information. This list is not just a theoretical idea; it is a coordinate-invariant map, meaning it stays the same no matter how you choose to label or measure the system.

To test this map, the team did not build new machines but instead re-examined data from three different public experiments that had already been performed. They treated these existing datasets as a laboratory to see if their new rules held up. In one experiment involving a logical Bell state created with neutral atoms, they analyzed 81 different settings of local measurements. They found that the information about the system's state was almost entirely concentrated in groups of three particles. Specifically, 99.719% of the measurable signal power came from groups of three, with almost nothing coming from smaller groups. This confirmed that for this specific setup, the information only becomes visible when you look at three particles at once. The data showed a clear plateau where the signal strength grew in a predictable way as the researchers adjusted the noise, matching their theoretical predictions for how information should behave under local disturbances.

The team also looked at data from a superconducting quantum system that used a dual-rail design, where information is carried by photons in two paths. Here, they tested how the system handled "erasure," a situation where a particle is lost or flagged as missing. Their analysis showed that if scientists only counted the successful attempts and ignored the lost ones, they would overestimate the amount of information available by about 16% to 21%. By including the information carried by the "flags" that marked the lost particles, they got a complete and accurate picture. This confirmed that the mathematical rules for counting information in the presence of noise were correct. Furthermore, they examined over 10 million outcomes from repeated gate operations to see if errors in one part of the system were linked to errors in another. They found that while the errors were mostly independent, there was a tiny, measurable connection between them that aligned with the direction of the information being measured.

These findings establish that the structure of quantum information is not random but follows strict rules about where it lives and how it reacts to noise. The researchers demonstrated that by looking at the "spectrum" of locality, one can predict exactly how a system will respond to different types of interference. They showed that for certain types of noise, the information loss follows a precise mathematical order, and for others, it follows a different, equally predictable pattern. The work does not claim to have solved all the mysteries of quantum sensing, but it provides a reproducible route to understanding them. It turns a vague question about "how much information is there" into a concrete answer about "where is it, and how big a piece do I need to see it?" By validating these rules against real-world data, the study offers a reliable guide for future experiments, ensuring that when scientists build the next generation of quantum devices, they know exactly how much of the system they need to watch to see the truth.

Technical Summary: Locality Spectroscopy of Quantum Statistical Information

Problem Statement
In multiparameter quantum estimation, a fundamental structural question arises: at what subsystem size does information about an unknown parameter first become locally identifiable? While prior research has quantified the total amount of Fisher information available, it has not systematically addressed the "structural" question of when independent statistical directions transition from being invisible to small subsystems to becoming visible globally. This is distinct from simply measuring the magnitude of information; it concerns the geometry of information distribution across a composite system.

Methodology
The authors define a coordinate-invariant locality spectrum derived from the rank births of reduced statistical tangents. For a differentiable multiparameter state family $\rho_\theta$ with tangents $\Delta_\mu = \partial_{\theta_\mu}\rho|_{\theta^\star}$ , the method involves:

Decomposition: Expressing tangents in a tensor-product orthonormal operator basis $\{B_\alpha\}$ and grouping coefficients by their support size (weight) $w_\alpha$ .
Spectrum Definition: Defining $R_k$ as the dimension of the span of coefficients with weight $\le k$ . The locality spectrum is the non-decreasing list $(d_1, \dots, d_r)$ , where $d_j = \min\{k : R_k \ge j\}$ . This spectrum is proven to be invariant under non-singular reparameterization.
Operational Link: Under the assumption that reduced baseline states are full rank, the spectrum is operationally linked to the rank of the sum of reduced Symmetric-Logarithmic-Derivative (SLD) Fisher matrices over all subsystems of size $k$ .
Noise Valuation: The authors derive analytical laws for how this spectrum manifests under specific local noise channels:
- Controlled Local Reset: For a product regular reset satisfying support-filtration injectivity, the order of the $j$ -th elementary symmetric Fisher volume vanishes as $2\alpha \sum_{i \le j} d_i$ .
- Flagged Erasure: For independent site retention with an observed erasure flag, the order vanishes as $\sum_{i \le j} d_i$ .
- Maximal Mixing: At a maximally mixed operating point, local depolarization and flagged erasure obey an exact Fisher-matrix substitution identity: $F_{dep}(\eta) = F_{era}(\eta^2)$ .

Key Results and Empirical Validation
The paper validates these theoretical constructs through the reanalysis of three distinct public datasets, separating exact source measurements from the authors' derived analyses:

Logical-Bell Experiment (Neutral Atoms):
- Data: 168,386 attempted shots from a [[4, 2, 2]] code experiment (Chung et al.).
- Finding: Direct Pauli-tangent analysis of the state $|\Psi_\phi\rangle = (|00\rangle_L + e^{i\phi}|11\rangle_L)/\sqrt{2}$ reveals that 99.719% of the debiased tangent power resides at Pauli weight 3. This confirms a strict single-parameter locality of $d=3$ within a finite practical window, despite minor lower-weight leakage detected by a joint Wald test.
- Noise Valuation: Analysis of reduced Quantum Fisher Information (QFI) sums ( $S_k$ ) shows a strong weight-3 hierarchy. When subjected to local depolarization and reset models, the effective exponents $d \ln F / d \ln x$ align with the predicted values of 3 and 6, respectively, within a broad finite-window plateau.
Dual-Rail Ramsey Data (Superconducting Qubits):
- Data: 825,000 attempts from Chou et al. involving phase sweeps and erasure flags.
- Finding: The study verifies "flagged-Fisher accounting." It demonstrates that discarding the erasure branch and reporting survivor-only Classical Fisher Information (CFI) overstates the information per successful preparation by a factor of 1.162–1.209. The explicit information contained in the erasure flag itself is negligible (<0.064% of full CFI), confirming the necessity of including the flag term for exact accounting.
Repeated-CZ Outcomes (Superconducting Qubits):
- Data: 10.8 million counts from repeated CZ gate experiments.
- Finding: The study separates erasure-flag correlations from correlations in surviving noise. While the erasure flags are nearly independent (mutual information $< 4.82 \times 10^{-5}$ bits), the surviving noise exhibits correlated dephasing.
- Metrological Impact: Despite the weak correlation in flags, the correlated noise sector (specifically $ZZ$ dephasing) yields a tangent-aligned phase-QFI ratio ( $F_{corr}/F_{prod}$ ) of 1.0076–1.0618. This indicates that nearly independent erasure statistics do not imply a product channel, and correlated noise remains metrologically relevant if aligned with the tangent.

Significance and Claims
The paper establishes locality spectroscopy as a reproducible route connecting subsystem identifiability to controlled-noise response. Its primary contributions are:

Formulating a structural object (the locality spectrum) that defines at what subsystem size independent statistical directions become visible, applicable to arbitrary differentiable multiparameter families.
Deriving exact noise-valuation laws for the full Fisher geometry, specifically linking the order of Fisher eigenvalue decay to the locality spectrum under reset and erasure channels.
Providing complementary experimental tests using public data to verify these laws, demonstrating that real-world quantum states often exhibit a "finite-window" adherence to ideal locality spectra rather than strict asymptotic behavior.

The authors explicitly note that no single existing source implements a purpose-built experiment with both a calibrated physical local-reset sweep and independent physical flagged erasure on the same state family; thus, the current work relies on reanalysis of separate datasets to test the theoretical framework. The paper concludes that these results provide a foundation for understanding how information geometry dictates the response of quantum systems to local noise.

Technical Summary: Locality Spectroscopy of Quantum Statistical Information

More like this