---
title: Randomized Measurements in Quantum Systems
url: https://www.emergentmind.com/topics/randomized-measurements
type: topic
---

# Randomized Measurements in Quantum Systems

Searching arXiv for recent and foundational papers on randomized measurements.
arxiv_search(query="all:\"randomized measurements\" OR all:\"classical shadows\" quantum", max_results=10, sort_by="submittedDate")
arxiv_search({"query":"all:\"randomized measurements\" OR all:\"classical shadows\" quantum","max_results":10,"sort_by":"submittedDate"})
Randomized measurements are a family of protocols for extracting properties of a quantum state or process by preparing the state, applying a random unitary \(U_{\mathrm{set}}\), measuring in the computational basis, and repeating over many random settings and shots. The same raw data can then be reused to estimate expectation values of many observables, purities, Rényi entropies, fidelities, and related quantities through classical post-processing, most prominently via the classical shadow formalism [2509.12749]. In many-body settings, the same paradigm also appears in random local Pauli measurements, collective random rotations, and observable-driven schemes tailored to specific nonlinear functionals such as \(\mathrm{Tr}(O\rho^2)\) [2411.03285].

## 1. Protocol and data model

In its standard form, a randomized-measurement protocol consists of four steps: prepare the state \(\rho\), apply a random unitary \(U_{\mathrm{set}}\), measure in the computational basis to obtain a bitstring \(\mathbf{s}\in\{0,1\}^N\), and repeat for many random unitaries and many shots per unitary [2509.12749]. A single measurement setting is often a local random unitary,
\[
U_{\mathrm{set}} = U_1\otimes\dots\otimes U_N,
\]
with single-qubit unitaries \(U_i\) drawn i.i.d. from an ensemble such as Haar on \(U(2)\) or the single-qubit Clifford group. Sampling \(N_{\mathrm U}\) such settings and repeating each \(N_{\mathrm M}\) times yields a dataset of \(N_{\mathrm U}\times N_{\mathrm M}\) bitstrings together with the list of unitaries used [2509.12749].

A particularly important specialization is the independent single-qubit randomized Pauli scheme. For an \(N\)-qubit state \(\rho\), each shot chooses \(P_i\in\{X,Y,Z\}\) independently and uniformly, so that
\[
p(P_1,\dots,P_N)=3^{-N},
\]
then measures each qubit in the eigenbasis of \(P_i\), producing outcomes \(b_i=\pm 1\) with conditional distribution
\[
p(b|P,\rho)=\mathrm{Tr}\Big(\rho \bigotimes_{i=1}^N \tfrac{1}{2}(1+b_iP_i)\Big).
\]
The pair \((P,b)\) is stored as one classical-shadow snapshot, and repeated shots give i.i.d. samples from \(p(P)p(b|P,\rho)\) [2411.03285].

Randomized measurements can also be implemented without active basis switching. A metasurface-based photonic protocol realizes a six-outcome POVM corresponding to the three Pauli bases \(x,y,z\), with effects \(\frac{1}{3}|\alpha\rangle\langle\alpha|\) for \(\alpha=x^\pm,y^\pm,z^\pm\). This fixed POVM is equivalent to choosing a random local Pauli basis and measuring projectively in it; mathematically it forms a single-qubit quantum 2-design [2308.08755].

## 2. Classical shadows and statistical reconstruction

The classical-shadow framework turns randomized-measurement data into random matrix estimators \(\hat\rho\) whose expectation equals the underlying state. For local unitary 3-designs such as local Cliffords, a classical snapshot is
\[
\hat{\rho} \;=\; \bigotimes_{i=1}^N \Bigl( 3\,U_i^\dag\ket{s_i}\!\bra{s_i}U_i - \openone \Bigr),
\]
and satisfies the unbiasedness relation \(\mathbb{E}[\hat\rho]=\rho\) [2509.12749]. In the randomized Pauli scheme, the inversion takes the product form
\[
\hat{\rho}(P,b)=\bigotimes_{i=1}^N \frac{1+3b_iP_i}{2},
\qquad
\mathbb{E}_{(P,b)}[\hat{\rho}(P,b)]=\rho,
\]
which makes the bridge between raw measurement data and physical observables completely explicit [2411.03285].

Given \(M\) independent snapshots \(\hat\rho_1,\dots,\hat\rho_M\), expectation values are estimated by
\[
\widehat{\langle O\rangle}=\frac{1}{M}\sum_{j=1}^M \operatorname{tr}(O\hat\rho_j),
\]
and the estimator is unbiased because \(\mathbb{E}[\hat\rho_j]=\rho\) [2509.12749]. In Pauli form, some observables can even be read off directly from \((P,b)\) without explicitly constructing \(\hat\rho\); for example,
\[
\langle Z_iZ_j\rangle
=
\mathbb{E}_{(P,b)}\big[b_ib_j\,\delta_{P_i=Z}\,\delta_{P_j=Z}\big],
\]
illustrating the “measure first, ask questions later” character of the data [2411.03285].

Nonlinear functionals require higher-order statistics. For trace moments \(p_k=\operatorname{tr}(\rho^k)\), unbiased estimators can be written as U-statistics over independent snapshots; these underlie estimators of Rényi entropies, mutual information, and mixed-state entanglement criteria [2509.12749]. Long before the language of classical shadows became standard, randomized local unitaries were already used to estimate subsystem purities through weighted correlations of measurement probabilities. For a subsystem \(A\) of \(N_A\) qudits of local dimension \(d\),
\[
\mathrm{Tr}(\rho_A^2)
=
d^{N_A}\sum_{\mathbf{s}_A,\widetilde{\mathbf{s}}_A}
(-d)^{-D[\mathbf{s}_A,\widetilde{\mathbf{s}}_A]}
\,\overline{P(\mathbf{s}_A)P(\widetilde{\mathbf{s}}_A)},
\]
which yields the second Rényi entropy \(S^{(2)}(\rho_A)=-\log_2\mathrm{Tr}(\rho_A^2)\) from single-copy randomized measurements [1806.05747].

## 3. Ensemble design and protocol variants

The baseline ensemble is the product ensemble of local random unitaries, typically drawn from single-qubit 3-designs such as local Clifford or Haar [2509.12749]. However, several variants modify the ensemble to target different observables, hardware constraints, or noise models.

Shallow shadows replace product unitaries by shallow random circuits. In that case the measurement channel is no longer analytically simple, so the effective map \(\mathcal M\) is learned from calibration data and inverted numerically, yielding unbiased estimators with significantly reduced variance for non-local observables [2509.12749]. Robust shadows incorporate calibration against time-stationary, Markovian, gate-independent measurement noise that twirls to a depolarizing form under randomization; the shadow map is then modified to invert both the ideal measurement channel and the effective depolarizing noise [2509.12749].

Randomized measurements can also be constrained to real subgroups. Real randomized measurements (RRMs) use orthogonal evolution and real local observables, while partial real randomized measurements (PRRMs) use orthogonal evolution and imaginary local observables. The two schemes capture different sectors of the bipartite correlation matrix: RRMs extract real–real correlations, PRRMs extract imaginary–imaginary correlations, and for real states their information coincides with that of standard randomized measurements [2411.06013].

A distinct adaptation arises in collective randomized measurements. Instead of independent local unitaries, one applies the same single-qubit unitary \(U\) to all particles, i.e. \(U^{\otimes N}\), and measures collective angular momentum such as \(J_z=\frac12\sum_i \sigma_z^{(i)}\). Haar moments of the random variable
\[
f_U(\rho)=\alpha\,J_{z,U}+\beta\,J_{z,U}^2+\gamma
\]
define collective randomized moments \(\mathcal J^{(r)}(\rho)=\int dU\,[f_U(\rho)]^r\), which are invariant under collective rotations and therefore collective-reference-frame–independent [2309.10745].

Randomization can also be used to simplify measurement noise itself. Randomized compiling for subsystem measurements inserts random Weyl or Clifford operations before and after a noisy computational-basis measurement, transforming a general noisy subsystem measurement into a uniform stochastic instrument whose effective description is a classical confusion matrix on the measured subsystem together with a stochastic channel on the unmeasured subsystem [2304.06599].

## 4. Entanglement, Rényi entropies, and correlation structure

Randomized measurements became prominent as a tool for entanglement diagnostics because they provide direct access to nonlinear quantities such as purities and Rényi entropies without full tomography or two-copy SWAP operations. In a trapped-ion quantum simulator with up to \(20\) ions, randomized local unitaries were used to measure second Rényi entropies of subsystems up to \(10\) qubits, to certify bipartite entanglement across all \(2^9-1=511\) distinct bipartitions of a \(10\)-qubit system, and to track the suppression of entanglement growth by disorder [1806.05747].

Beyond purity estimation, moments of randomized correlation distributions furnish local-unitary-invariant diagnostics of multipartite structure. For \(n\)-qubit states, second moments of random correlations of the full system and its marginals can be combined to obtain the global purity,
\[
\operatorname{tr}(\varrho^2)=\frac{1}{2^n}\sum_{\mathcal A\subseteq\mathcal S}3^{|\mathcal A|}m_{\mathcal A},
\]
while suitable regions in the \((\mathcal R^{(2)},\mathcal R^{(4)})\) plane distinguish biseparable from genuinely multipartite entangled states and exclude membership in mixed \(W\)-class states [2011.10591].

The same logic extends to weakly entangled and bound-entangled states. For three qubits, optimal linear criteria in the sector-length variables \(A_1,A_2,A_3\) were derived from second moments, and for higher-dimensional bipartite systems fourth moments of randomized measurements detect PPT bound entangled states such as chessboard, UPB-based, Horodecki, and \(4\otimes4\) Piani states [2010.08372]. This establishes that randomized measurements are not restricted to purity-like witnesses, but can access genuinely higher-order correlation structure.

Collective randomized measurements provide an alternative route when only collective control is available. For permutationally symmetric states, the first three collective randomized moments \(\mathcal J^{(r)}\) determine the eigenvalues of the two-qubit covariance matrix \(C\), thereby completely characterizing spin-squeezing entanglement. With a different choice of moments, the criterion
\[
\sum_{l=x,y,z}(\Delta J_l)^2 < \frac{N}{2}
\]
is recovered in collective-reference-frame–independent form, and a further antisymmetric three-body criterion detects entanglement, including bound entanglement, beyond optimal spin-squeezing inequalities [2309.10745].

## 5. Metrology, nonlinear functionals, and algorithmic uses

Quantum Fisher information (QFI) is a particularly important nonlinear functional. One route expresses QFI through a monotonically increasing family of polynomial lower bounds \(F_n(\rho,A)\), with \(F_n\le F_{n+1}\le F_Q\), and estimates these polynomials from classical shadows via U-statistics. The first bounds,
\[
F_0 = 4\,\mathrm{Tr}(\rho[\rho,A]A),
\qquad
F_1 = 2F_0 - 4\,\mathrm{Tr}(\rho^2[\rho,A]A),
\]
can be estimated from randomized measurements without full tomography and used as multipartite-entanglement witnesses through the standard QFI-depth criteria [2105.13164].

A complementary experimental route estimates the superfidelity between \(\rho_\theta\) and \(\rho_{\theta+d\theta}\) from randomized measurements, defines the sub-quantum Fisher information \(\mathcal F_G\), and uses \(\mathcal F_\theta\ge \mathcal F_G\). On an NV-center spin in diamond, the method yields the exact QFI; on a noisy 4-qubit GHZ-type state on an IBM superconducting processor it gave \(\mathcal F_G^{\text{rand}}=4.301\pm 0.22\), compared with \(\mathcal F_G^{\text{tomo}}=4.4807\) and exact \(\mathcal F=6.5147\) from tomography [2104.00519].

Recent work has generalized the randomized-measurement toolbox from purity to arbitrary nonlinear functionals of the form \(\mathrm{Tr}(O\rho^2)\). Observable-driven randomized measurement (ORM) decomposes \(O\) into dichotomic observables, applies block-diagonal random unitaries adapted to the eigenspaces of \(O\), and estimates \(\mathrm{Tr}(O\rho^2)\) with sample complexity
\[
O\!\left(
\max\left\{\frac{\sqrt d}{\varepsilon},\frac{1}{\varepsilon^2}\right\}
\log\!\left[\frac{1}{\delta}\log\frac{1}{\varepsilon}\right]
\right)
\]
for \(\|O\|_\infty\le 1\). For all Pauli observables, this scaling is optimal up to logarithmic factors, and a related braiding randomized measurement protocol estimates low-rank observables such as projectors relevant to fidelity-like quantities [2505.09206].

Randomized measurements also support algorithmic decompositions of quantum computations. In fast circuit cutting, the identity channel on \(k\) wires is decomposed into randomized measure-and-prepare channels, yielding an unbiased estimator of the original circuit output with sampling overhead \(\widetilde O(4^k/\varepsilon^2)\) and an information-theoretic lower bound \(\Omega(2^k/\varepsilon^2)\) for any comparable procedure. Applied to QAOA, this gives simulation on smaller devices with overhead roughly \(2^{O(p\kappa)}\), where \(p\) is the number of entangling layers and \(\kappa\) is the size of a balanced vertex separator of the problem graph [2207.14734].

In many-body learning, randomized measurements can serve as training data rather than only post-processing input. ShadowGPT is trained on simulated classical shadow data of ground states generated by randomized Pauli measurements and learns the conditional distribution \(p(b|P,g)\) over outcomes \(b\) given Pauli strings \(P\) and Hamiltonian parameters \(g\). Once trained, it generates synthetic shadows from which ground-state energy, correlation functions, and second Rényi entanglement entropy are estimated for new parameters in the transverse-field Ising model and the \(\mathbb Z_2\times\mathbb Z_2\) cluster–Ising model [2411.03285].

## 6. Noise mitigation, hardware implementations, and software

Real devices motivate noise-aware randomized measurements. A self-calibrating framework connects randomized benchmarking and shadow estimation through a non-commutative Fourier transform of the noisy implementation map. For gate-independent noise, randomized-benchmarking decay parameters determine the noisy frame operator, allowing a calibrated inverse map and robust shadow estimation. In a CNOT-dihedral plus Clifford scheme, the same randomized-measurement data can be used both to estimate the decay parameter \(\lambda_Z\) and to build noise-mitigated shadows, so calibration and shadow estimation occur in a single experimental session [2403.04751].

On the hardware side, metasurface-based randomized measurements realize a fixed photonic POVM forming a single-qubit quantum 2-design, with six output ports corresponding to the Pauli-\(X\), \(Y\), and \(Z\) eigenstates. The induced measurement channel obeys
\[
\hat M(\rho)=\frac{I+\rho}{3},
\qquad
\hat M^{-1}(X)=3X-I,
\]
exactly as in Pauli classical shadows. A calibrated noise model including basis-flip, amplitude-damping–like errors, and basis- and outcome-dependent photon loss is inverted at the probability level, and the resulting error-mitigated estimators recover fidelities and purities in simulated W-state benchmarks [2308.08755].

Software has matured around the full randomized-measurement workflow. “RandomMeas.jl” implements pre-processing through `MeasurementSetting` types, data acquisition into `MeasurementData` and `MeasurementGroup`, and post-processing into `FactorizedShadow`, `DenseShadow`, and `ShallowShadow`. It supports expectation values, trace moments, Rényi entropies, purity and overlap estimation, XEB and self-XEB, jackknife uncertainty estimates, and tensor-network-based simulation through ITensors.jl [2509.12749].

## 7. Limitations, scaling, and emerging directions

Despite their breadth, randomized measurements retain nontrivial statistical bottlenecks. In purity estimation with local Haar-random unitaries, the integrand
\[
X(u)=2^N\sum_{s,s'}(-2)^{-D[s,s']}P_u(s)P_u(s')
\]
can vary over the full range \(2^{-N}\le X(u)\le 2^N\), so uniform Monte Carlo is inefficient and the required number of total measurements typically scales as \(N_uN_M\sim 2^{aN}\) with \(a\approx 1\) [2102.13524]. Importance sampling improves this markedly: for product states the paper derives \(a\approx 0.37\) in a moderate-size regime, confirms exponential savings numerically for product and GHZ states, and shows consistent gains for random and quenched states [2102.13524].

Other limitations are estimator-specific. Local Pauli shadow schemes degrade for high-weight nonlocal observables: in ShadowGPT, long \(X\)-string operators are predicted visibly worse than local correlators such as \(Z_iZ_{i+n}\), consistent with the large variance of Pauli-based shadow estimators for high-weight operators [2411.03285]. Tensor-network simulation of randomized measurements is efficient only for states with limited entanglement, such as 1D systems with modest bond dimension or shallow circuits; highly entangled dynamics may be out of reach [2509.12749]. Robust-shadow analyses also assume noise models that are Markovian, stationary, and gate-independent enough to twirl to simple effective channels [2403.04751].

Current developments point in several directions. This suggests a convergence between randomized measurements, machine learning, and adaptive sampling. Observable-driven schemes extend optimal single-copy estimation beyond purity to \(\mathrm{Tr}(O\rho^2)\) [2505.09206]. Real and partial real randomized measurements separate real and imaginary correlation sectors and yield lower bounds on the robustness of imaginarity [2411.06013]. Self-calibrating shadows suggest that benchmarking data can inform learning procedures directly [2403.04751]. Software roadmaps include process shadows, symmetry-resolved shadows, optimized sampling and derandomization, and GPU-enabled tensor-network backends [2509.12749]. Together these directions indicate that randomized measurements have become not merely a family of estimators, but a general interface between quantum experiments, nonlinear state functionals, and classical inference.

Source: https://www.emergentmind.com/topics/randomized-measurements