Nonadaptive Pauli Measurements

• Nonadaptive Pauli measurements are protocols with pre-determined settings for projective measurements in the Pauli basis, ensuring statistical independence across quantum state copies.
• They enable scalable quantum state and process tomography by demonstrating provable near-optimal sample complexity in various estimation tasks.
• Their simplicity and compatibility with current hardware make these measurements ideal for efficient error correction, resource certification, and compressed sensing applications.

Nonadaptive Pauli measurements denote protocols in which the choice of measurement settings—namely, projections onto Pauli observables (tensor products of $I$, $X$, $Y$, $Z$)—is fixed in advance, with outcomes collected from independently prepared copies of a quantum state or channel. No intermediate feed-forward, classical feedback, or reconfiguration based on prior measurement outcomes occurs. This paradigm spans key applications in quantum state and process tomography, fidelity estimation, channel verification, resource certification, and quantum error correction. Nonadaptive Pauli measurement strategies are favored for their implementation simplicity, compatibility with current hardware, support for rigorous theoretical analysis, and—in several regimes—provable near-optimality.

1. Operational Definition and Measurement Model

A nonadaptive Pauli measurement protocol specifies in advance a list or distribution over Pauli observables to be measured. On each independent quantum system copy, one performs the associated projective measurement (or more generally a POVM with elements diagonal in the Pauli basis) and records the resulting outcomes. For the $n$-qubit setting, the relevant Pauli group is $$\mathcal{P}_n = \{P = P_1 \otimes \cdots \otimes P_n : P_k \in \{I,X,Y,Z\}\}$$, supporting up to $4^n$ distinct observables.

In state tomography, certification, or reduction tasks, the nonadaptive feature manifests as statistical independence: each measurement outcome is determined only by the selected observable and the underlying system state, and the complete measurement record is analyzed collectively post hoc. Prominent examples include:

• Certification of Clifford-enhanced product states (e.g., magic-state injection outputs): Each copy is subjected to a single-qubit Pauli measurement on one qubit, with observable chosen from a fixed importance-weighted distribution determined by the target state (Sater et al., 10 Nov 2025).
• Full or partial tomography: The entire set, or a well-chosen subset, of multi-qubit Pauli operators is measured independently to reconstruct the quantum state (Acharya et al., 25 Feb 2025, Yu, 2020).
• Process characterization: Pauli measurements executed as part of entanglement-assisted tomography or channel learning, where the channel is applied to one half of a maximally entangled pair and Bell-basis (which are Pauli-diagonal) measurements are performed (Trinh et al., 2024).

Nonadaptivity is distinguished from adaptive measurement schemes, in which subsequent measurements depend on prior outcomes, as well as from collective protocols which entangle or jointly process multiple system copies.

2. Sample Complexity and Near-Optimality

The sample (copy) complexity of nonadaptive Pauli measurement protocols is sharply characterized for several central estimation tasks:

Quantum state tomography:

• For arbitrary $n$-qubit states, nonadaptive Pauli tomography achieves infidelity $\varepsilon$ with $O(10^n/\varepsilon^2)$ copies (Acharya et al., 25 Feb 2025, Acharya et al., 29 Jul 2025, Yu, 2020).
• Information-theoretic lower bounds show that even adaptive Pauli-only schemes require $\Omega(9.118^n/\varepsilon^2)$ copies, providing the first explicit separation from structured POVMs (e.g., SIC, MUB, uniform) which can attain the optimal $\Theta(8^n/\varepsilon^2)$ scaling. Thus, Pauli-based approaches are strictly sub-optimal by a constant-factor in the exponential base (Acharya et al., 25 Feb 2025).
• For single-qubit (product) measurements, bounds are tight up to a subpolynomial (e.g., $\sqrt{n}$) factor: $O(10^n/\varepsilon^2)$ upper bound vs. $\Omega(10^n/(\sqrt{n} \varepsilon^2))$ lower bound (Acharya et al., 29 Jul 2025).
• For pure-state tomography, recent advances show nonadaptive Pauli measurements can attain $N = \widetilde{O}(2^n/\varepsilon)$ copies, matching the optimal scaling achieved by general measurements up to polylogarithmic factors (Grewal et al., 7 Jan 2026).
• Minimal nonadaptive, informationally-complete Pauli sets for pure states have been constructed for low $n$ ($11$ for $n=2$, $31$ for $n=3$), showing significant reduction compared to full tomography (Ma et al., 2016).

Quantum channel learning/Pauli channel identification:

• For $n$-qubit Pauli channels, nonadaptive Bell-pair–assisted Pauli measurements are strictly optimal for learning and hypothesis testing in any $\ell_p$ norm, with query complexity scaling as $\Theta(4^n/\varepsilon^2)$ for $\ell_1$ distance, and as $\Theta(4^n/(n\varepsilon^2))$ for entropy or support estimation (Trinh et al., 2024).

Resource-efficient tomography (compressed sensing):

• For rank-$r$ quantum states, nonadaptive Pauli measurements furnish universal recovery with $O(r d \mathrm{polylog}(d))$ randomly chosen observables (where $d=2^n$), provided nuclear-norm minimization is used for reconstruction. This leverages the restricted isometry property (RIP) for rank-$r$ matrices (Liu, 2011).

3. Protocol Construction and Statistical Estimators

Nonadaptive Pauli measurement protocols rely on essential statistical estimators built from measurement outcomes:

• Full Pauli tomography: Each Pauli $P$ is measured $m$ times to estimate $\mathrm{Tr}(\rho P)$, and the state estimate is reconstructed as $$\widehat{\rho} = \frac{1}{d} \sum_P \widehat{\alpha}_P P$$, where $\widehat{\alpha}_P$ is the sample average over outcomes. The estimator is unbiased, with variance dominated by the inverse of the number of times each $P$ is measured, taking into account redundancy due to overlapping supports (Acharya et al., 25 Feb 2025, Yu, 2020).
• Importance sampling for direct fidelity estimation: For a target pure state $\rho$, importance sampling over Pauli observables according to $p_k = (\mathrm{Tr}[\rho W_k]/\sqrt{d})^2$ ensures that only a constant number of measurements are required to estimate fidelity up to constant additive error, yielding an exponential advantage over full tomography for high-fidelity states (Flammia et al., 2011).
• Certification of Clifford-enhanced product states: Each measurement is chosen from a distribution proportional to $| \langle \psi_i | P | \psi_i \rangle |$ across all single-qubit targets and axes, yielding a robust estimator for the global fidelity via signed outcomes; fidelity is “witnessed” via aggregation and compared to a threshold to decide acceptance (Sater et al., 10 Nov 2025).

Sample complexity is derived using concentration bounds (McDiarmid, Hoeffding, Chernoff), and tightness is ensured by information-theoretic techniques leveraging packing arguments, mutual information, and the eigenvalue structure of the measurement information channel.

4. Hardware Implementations and Grouping Strategies

Physical implementation of nonadaptive Pauli measurements leverages available hardware primitives:

• Multi-qubit projective measurement of Pauli stabilizers: Via dispersive circuit QED modules, arbitrary multi-qubit Pauli operators are projected nonadaptively—an ancilla is entangled with the code register through an engineered interaction, and measurement of the ancilla implements the projective measurement on the code (Blumoff et al., 2016). This hardware supports both single- and multi-qubit Pauli measurements required for quantum error correction and stabilizer readout.
• Measurement grouping and commuting collections: When estimating expectation values of observables expressible as sparse sums of Pauli terms (e.g., molecular Hamiltonians), grouping mutually commuting sets of Pauli observables minimizes the required measurement shots. Algorithms such as SORTED INSERTION maximize the state-independent improvement ratio $\hat R$, outperforming heuristics which simply minimize group count. Efficient Clifford circuits (CZ- or CNOT-based) implement the joint measurement of each commuting group, optimizing two-qubit gate counts (Crawford et al., 2019).

5. Certification, Robustness, and Extensions

Nonadaptive Pauli measurement protocols provide strong robustness and theoretical guarantees:

• Certification (soundness and completeness): For Clifford-enhanced product states, certification via single-qubit nonadaptive Pauli measurements achieves optimal soundness/completeness scaling in sample complexity, with precise thresholds distinguishing high-fidelity preparations from mispreparations (Sater et al., 10 Nov 2025).
• Adaptivity does not help: In the context of Pauli-based tomography and channel estimation, adaptivity yields at most constant, sub-exponential improvements; entangled or adaptive measurements offer no asymptotic advantage for a broad class of tasks, including Pauli channels and full quantum state tomography, as established by matching upper/lower bounds (Lowe et al., 2022, Trinh et al., 2024).
• Overlap and partial tomography: Quantum "overlapping tomography," in which one seeks to reconstruct all $k$-qubit marginals, can be optimally accomplished nonadaptively with $O(10^k \log(\binom{n}{k}/\delta)/\varepsilon^2)$ copies, and no entangled measurement protocol can outperform this scaling for fixed $k$ (Yu, 2020).
• Robustness to noise and errors: Nonadaptive schemes remain robust to moderate levels of depolarizing noise, producing high-fidelity reconstructions in simulated and experimental NMR platforms for low $n$ (Ma et al., 2016).

6. Applications and Research Directions

Nonadaptive Pauli measurement protocols are foundational in:

• Quantum state and process tomography (standard and compressed-sensing variants)
• Direct certification and benchmarking of quantum hardware (magic-state factories, Clifford circuits, etc.)
• Quantum error correction (syndrome extraction, stability measurement, fault-tolerant initialization)
• Variational quantum algorithms (Hamiltonian expectation estimation via optimized grouping)
• Experimental and resource-efficient protocols in NMR, optics, and superconducting platforms

Active research continues on the construction of minimal informationally-complete nonadaptive Pauli sets for larger $n$, closing remaining polylogarithmic sample complexity gaps, and refining classical post-processing for higher efficiency and robustness (Ma et al., 2016, Grewal et al., 7 Jan 2026).

---

Summary Table: Sample Complexity of Nonadaptive Pauli Measurements for Major Quantum Estimation Tasks

Task	Sample Complexity	Reference
Full arbitrary state tomography	$O(10^n/\varepsilon^2)$	(Acharya et al., 25 Feb 2025, Acharya et al., 29 Jul 2025, Yu, 2020)
Pure state tomography (best known)	$\widetilde O(2^n/\varepsilon)$	(Grewal et al., 7 Jan 2026)
Overlapping tomography (fixed $k$)	$O(10^k\,\log(\binom{n}{k}/\delta)/\varepsilon^2)$	(Yu, 2020)
Clifford-enhanced product-state certification	$O(n^2/\varepsilon^2)$	(Sater et al., 10 Nov 2025)
Pauli channel learning ($\ell_1$)	$\Theta(4^n/\varepsilon^2)$	(Trinh et al., 2024)
Low-rank ($r$) tomography (compressed sensing)	$O(rd\,\mathrm{polylog} d)$	(Liu, 2011)

The nonadaptive Pauli measurement paradigm offers a unifying and operationally tractable framework enabling efficient, scalable, and theoretically guaranteed certification and learning in quantum information science.