betaFeedback

WorksMathematics

Bayesian Weak Commutativity Does Not Guarantee Attainability of the Nagaoka–Hayashi Bound

Abstract

Bayesian weak commutativity has been suggested as the most likely condition for reaching the Nagaoka–Hayashi (NH) bound in multiparameter estimation. We show that it is not sufficient. For a full-rank four-level ensemble obeying it exactly, the effects required for NH equality cannot sum to the identity. We find the optimal risk exactly: it lies strictly above the bound and needs seven outcomes, while projective readout on the four levels adds over seven times the NH gap. A robust normalization witness bounds the gap from below across a family including truncated oscillator ladders.

Introduction—When a quantum probe must reveal several parameters at once, such as the components of a vector field [1], the optimal measurements for the individual parameters are generally incompatible [2, 3, 4, 5]. With collective measurements on many copies, the quantum Cramér–Rao bound is asymptotically attainable exactly when the symmetric logarithmic derivatives commute on average in the probe state, the weak-commutativity condition [4]. For single-copy, locally unbiased estimation, the Nagaoka–Hayashi (NH) bound extends Nagaoka’s two-parameter bound [6, 7] to more parameters [8, 9]; it is a semidefinite program [9] with a Bayesian counterpart [10]. In the local setting, a vanishing averaged commutator does not guarantee that NH is reached: a rank-three four-level example misses the bound [11].

Bayesian estimation asks a different question: the parameters follow a known prior, estimates may be biased, and one minimizes the prior-averaged squared error [10, 12, 13, 14]. Albarelli, Branford, and Rubio [14] showed that the Bayesian NH bound relaxes one requirement, namely that the first and second moments of the estimates come from a single normalized measurement. They proved that the bound is reached for two-parameter qubit models and suggested Bayesian weak commutativity as the most likely general condition for this; all their examples satisfy it and appear to reach the bound. Neither the local counterexample nor recent related results settle this: a different Bayesian bound attainable for qubits [15], numerical NH saturation for a qubit channel [16], and bounds on the number of outcomes an optimal measurement needs [17].

In this Letter, we show that the condition is not sufficient. A receiver is given one copy of a four-level system, prepared in one of four known full-rank states, and estimates two real coordinates; we minimize the prior-averaged squared error over all positive operator-valued measures (POVMs) and estimators. Weak commutativity holds exactly, but every effect required for NH equality has the same fixed ratio between two level populations, and the identity does not, so these effects can never form a complete measurement. A normalization witness, an observable that is nonpositive on all such effects but has positive trace, captures this mismatch, and its robust form makes the risk gap quantitative. We find the exact optimal risk and a seven-outcome measurement attaining it, and prove that fewer outcomes cannot. Projective measurements (PVMs) on the same four levels have a strictly larger exact minimum, attained with four outcomes.

The obstruction is not tied to one ladder: across a family that contains truncated oscillator and pair-annihilation ladders, a robust witness shows that at least 1/34201/3420 of the risk reduction predicted by NH is unattainable, and on the oscillator slice we find the optimal measurement exactly.

The estimation task—A prior π\pi selects a dimensionless label θ∈R2\theta\in\mathbb{R}^{2} and a state ρθ\rho_\theta on C4\mathbb{C}^{4}. For unbiased squared error, only three prior moments matter: the average state S=EπρθS = \mathbb{E}_{\pi}\rho_\theta, the label–state correlations Bj=Eπ(θjρθ)B_j = \mathbb{E}_{\pi}(\theta_j\rho_\theta), and λ=Eπ∥θ∥2\lambda= \mathbb{E}_{\pi}\lVert\theta\rVert^{2}, which is the risk of always reporting zero. A measurement is a normalized, countably additive POVM PP on an arbitrary outcome space O\mathcal{O}, followed by a measurable report a:O→R2a:\mathcal{O}\to\mathbb{R}^{2}. Its risk is

L(P,a)=Eπ∫O∥a(ω)−θ∥2Tr⁡[ρθP(dω)].(1)\mathcal{L}(P,a)=\mathbb{E}_{\pi}\int_{\mathcal{O}}\lVert a(\omega)-\theta\rVert^{2}\operatorname{Tr}[\rho_\theta P(\mathrm{d}\omega)]. \tag*{(1)}

We write Lmin⁡\mathcal{L}_{\min} for the infimum over all measurements and reports. The symmetric posterior-mean operators SjS_j are defined by SSj+SjS=2BjSS_j+S_jS=2B_j [12], and Bayesian weak commutativity is the condition

Tr⁡(S[S1,S2])=0.(2)\operatorname{Tr}(S[S_1,S_2])=0. \tag*{(2)}

It constrains only this average; the two operators themselves need not commute. For Hermitian V1,V2V_1,V_2, the two-parameter NH bound reads

LNH=λ+inf⁡V1,V2{Tr⁡S(V12+V22)−2∑j=12Tr⁡BjVj+∥iS1/2[V1,V2]S1/2∥1}.(3)\mathcal{L}_{\mathrm{NH}}=\lambda+\inf_{V_1,V_2}\left\{\operatorname{Tr}S(V_1^2+V_2^2)-2\sum_{j=1}^{2}\operatorname{Tr}B_jV_j+\left\lVert iS^{1/2}[V_1,V_2]S^{1/2}\right\rVert_1\right\}. \tag*{(3)}

The last term is a trace norm, and, unlike in the local bound, the pair V1,V2V_1,V_2 is not constrained by unbiasedness. Equation (3) and (2) are Eqs. (45) and (75) of Ref. [14], whose Sec. 6 suggests the latter as the most likely attainability condition.

A four-level counterexample—Let C=∣1⟩⟨2∣+2∣2⟩⟨3∣+3∣3⟩⟨4∣C = |1\rangle\langle2| + 2|2\rangle\langle3| + 3|3\rangle\langle4| be a weighted ladder operator, and let X=(C+C†)/2X = (C + C^\dagger)/2 and Y=(C−C†)/(2i)Y = (C - C^\dagger)/(2i) be its two finite-level quadratures. At the maximally mixed average S=I/4S = I/4 the trace in Eq. (2) vanishes for every ensemble, so we make the populations unequal. With the population ratio k=S33/S11=1000001/1000000k = S_{33}/S_{11} = 1000001/1000000, define

p=[2+2k−4(k−1)9(1+k)2]−1,q=kp,t=q−4p(k−1)9(1+k)2,(4)p = \left[2 + 2k - \frac{4(k-1)}{9(1+k)^2}\right]^{-1}, \qquad q = kp,\quad t = q - \frac{4p(k-1)}{9(1+k)^2}, \tag*{(4)}
Sk=diag⁡(p,p,q,t).S_k = \operatorname{diag}(p,p,q,t).

Unless stated otherwise, kk is fixed at this rational value. Set U=diag⁡(−1,−1,−1,1)U = \operatorname{diag}(-1,-1,-1,1), K±,k=Sk(I±U)/2K_{\pm,k} = S_k(I \pm U)/2, and Bk=CK+,k+K−,kCB_k = CK_{+,k} + K_{-,k}C. The nonzero entries of BkB_k, just above the diagonal, are β1=p\beta_1 = p, β2=2p\beta_2 = 2p, and β3=3(q+t)\beta_3 = 3(q+t); we write Bk=B1,k+iB2,kB_k = B_{1,k} + iB_{2,k} with Hermitian Bj,kB_{j,k}. The four equiprobable preparations are

θs,j=10sej,ρs,j=Sk+sBj,k/5,s=±1, j=1,2.(5)\theta_{s,j} = 10s e_j,\qquad\rho_{s,j} = S_k + sB_{j,k}/5,\qquad s=\pm1,\ j=1,2. \tag*{(5)}

They have average state SkS_k, first moments Bj,kB_{j,k}, and λ=100\lambda= 100. Diagonal dominance gives ρs,j⪰I4/20\rho_{s,j} \succeq I_4/20, so every preparation has full rank. The posterior-mean operators have upper entries hjh_j and −ihj-ih_j, with h1=1/2h_1 = 1/2, h2=2p/(p+q)h_2 = 2p/(p+q), and h3=3h_3 = 3. Thus

Tr⁡(Sk[S1,k,S2,k])=2i[(p−q)h22+9(q−t)]=0.(6)\operatorname{Tr}(S_k[S_{1,k},S_{2,k}]) = 2i[(p-q)h_2^2 + 9(q-t)] = 0. \tag*{(6)}

The choice of tt makes the two terms cancel; since the populations are unequal, this is a genuine constraint rather than an identity. We use the phase symmetry of the moments only to construct measurements, never to restrict competing ones.

This single task has three distinct precision levels:

LNH,k<Lmin⁡,k<LPVM,k.(7)\mathcal{L}_{\mathrm{NH},k} < \mathcal{L}_{\min,k} < \mathcal{L}_{\mathrm{PVM},k}. \tag*{(7)}

Here LPVM,k\mathcal{L}_{\mathrm{PVM},k} is the infimum over projective measurements on the original C4\mathbb{C}^4 with arbitrary reporting rules. The optimum over general measurements and its distance from NH are

Lmin⁡,k=λ−∑ici,k,Γk=Tr⁡T0,k−∑ici,k>1/32,(8)\mathcal{L}_{\min,k} = \lambda- \sum_i c_{i,k},\qquad\Gamma_k = \operatorname{Tr} T_{0,k} - \sum_i c_{i,k} > 1/32, \tag*{(8)}

where T0,k=diag⁡(0,p,4p+9t,9q)T_{0,k} = \operatorname{diag}(0,p,4p+9t,9q) and the algebraic numbers ci,kc_{i,k} are defined by the construction below and in Appendix D. Their certified enclosures give Lmin⁡,k≈94.2859248008006880\mathcal{L}_{\min,k} \approx94.2859248008006880 and Γk≈0.0359263355222758\Gamma_k \approx0.0359263355222758, and the exact projective minimum of Eq. (15) is LPVM,k≈94.5557442485432417\mathcal{L}_{\mathrm{PVM},k} \approx94.5557442485432417. The gap Γk\Gamma_k is thus about 0.6%0.6\% of the risk reduction predicted by NH, λ−LNH,k=Tr⁡T0,k≈5.75\lambda-\mathcal{L}_{\mathrm{NH},k} = \operatorname{Tr} T_{0,k} \approx5.75, whereas the extra cost of projective readout, LPVM,k−Lmin⁡,k≈0.270\mathcal{L}_{\mathrm{PVM},k}-\mathcal{L}_{\min,k} \approx0.270, is more than seven times

larger Figure 1(b). Table 1 in Appendix E summarizes the three benchmarks.

NH-equality effects and Bayes risk comparison

Figure 1. Why weak commutativity is not enough, for the ensemble of Eq. (5). (a) Two diagonal entries of an effect EE. Every effect allowed by NH equality, and every sum of them, lies on the ray 9E22=4E339E_{22}=4E_{33}, where Tr⁡WE=0\operatorname{Tr}WE=0 for the witness W=diag⁡(0,9,−4,0)W=\operatorname{diag}(0,9,-4,0) (arrow, normal to the ray). The identity lies in the shaded region Tr⁡WE>0\operatorname{Tr}WE>0 ( Tr⁡W=5\operatorname{Tr}W=5) and cannot be reached. (b) NH bound (dashed) and exact optima over all POVMs and over PVMs on C4\mathbb{C}^4, rounded, on a truncated axis (no error bars); arrows mark Γk\Gamma_k, and the projective excess. Generative AI assisted the figure design and TikZ source.

BenchmarkEstablished result
NH relaxationStrictly below achievable risk, gap ≈0.036\approx0.036
General POVMExact minimum; seven outcomes are minimal
Original-space PVMExact minimum, four outcomes; excess ≈0.270\approx0.270, analytically >25/192>25/192

Table 1. Precision benchmarks for the four-state prior in Eq. (5), with λ=100\lambda=100 and dimensionless squared error. The PVM excess is relative to the general optimum and includes parameter-independent mixtures of fixed projective settings on the original C4\mathbb{C}^4. Projective implementations of POVMs on larger spaces are not restricted.

References

  1. [1]R. Kaubruegger, A. Shankar, D. V. Vasilyev, and P. Zoller, Optimal and variational multiparameter quantum metrology and vector-field sensing, PRX Quantum 4, 020333 (2023), arXiv:2302.07785.DOI
  2. [2]C. W. Helstrom, Quantum Detection and Estimation Theory (Academic Press, New York, 1976).
  3. [3]A. S. Holevo, Probabilistic and Statistical Aspects of Quantum Theory, 2nd ed. (Edizioni della Normale, Pisa, 2011).
  4. [4]S. Ragy, M. Jarzyna, and R. Demkowicz-Dobrzański, Compatibility in multiparameter quantum metrology, Phys. Rev. A 94, 052108 (2016), arXiv:1608.02634.
  5. [5]R. Demkowicz-Dobrzański, W. Górecki, and M. Guţă, Multi-parameter estimation beyond quantum Fisher information, J. Phys. A 53, 363001 (2020), arXiv:2001.11742.
  6. [6]H. Nagaoka, A new approach to Cramér–Rao bounds for quantum state estimation, in Asymptotic Theory of Quantum Statistical Inference: Selected Papers, edited by M. Hayashi (World Scientific, Singapore, 2005), pp. 100–112, reprinted from IEICE Tech. Rep. IT89-42 (1989).DOI
  7. [7]H. Nagaoka, A generalization of the simultaneous diagonalization of Hermitian matrices and its relation to quantum estimation theory, Trans. Jpn. Soc. Ind. Appl. Math. 1, 305 (1991), in Japanese.
  8. [8]M. Hayashi, On simultaneous measurement of noncommutative observables, in Development of Infinite-Dimensional Noncommutative Analysis, RIMS Kôkyûroku, Vol. 1099, edited by A. Hora (Research Institute for Mathematical Sciences, Kyoto University, Kyoto, 1999) pp. 96–118, in Japanese.
  9. [9]L. O. Conlon, J. Suzuki, P. K. Lam, and S. M. Assad, Efficient computation of the Nagaoka–Hayashi bound for multiparameter estimation with separable measurements, npj Quantum Inf. 7, 110 (2021).DOI
  10. [10]J. Suzuki, Bayesian Nagaoka–Hayashi bound for multiparameter quantum-state estimation problem, IEICE Trans. Fundam. Electron. Commun. Comput. Sci. E107.A, 510 (2024).arxiv.org/abs/2302.14223
  11. [11]L. O. Conlon, J. Suzuki, P. K. Lam, and S. M. Assad, Role of the extended Hilbert space in the attainability of the quantum Cramér–Rao bound for multiparameter estimation, Phys. Lett. A 542, 304445 (2025), arXiv:2404.01520.arxiv.org/abs/2404.01520
  12. [12]S. Personick, Application of quantum estimation theory to analog communication over quantum channels, IEEE Trans. Inf. Theory 17, 160 (1971).DOI
  13. [13]J. Rubio and J. Dunningham, Bayesian multiparameter quantum metrology with limited data, Phys. Rev. A 101, 032114 (2020), arXiv:1906.04123.DOI
  14. [14]F. Albarelli, D. Branford, and J. Rubio, Measurement incompatibility in Bayesian multiparameter quantum estimation, Quantum 10, 2192 (2026), arXiv:2511.16645.DOI
  15. [15]K.-H. Zhao, K. Yamagata, and J. Suzuki, Bayesian Gill–Massar bound: An attainable lower bound for qubit parameter estimation (2026), arXiv:2607.07031.arxiv.org/abs/2607.07031
  16. [16]L. Bia and C. N. Gagatsos, Saturating the Bayesian Nagaoka–Hayashi bound within numerical precision for the depolarization SU(2) rotation channel (2026), Phys. Rev. A (to be published), doi:10.1103/p7pv-vyxf, arXiv:2607.15398.arxiv.org/abs/2607.15398
  17. [17]K. Yamagata, Sufficient support size of measurements for quantum estimation (2026), arXiv:2604.21323.arxiv.org/abs/2604.21323
  18. [18]A. S. Holevo, Statistical decision theory for quantum systems, J. Multivariate Anal. 3, 337 (1973).DOI
  19. [19]M. Hayashi and Y. Ouyang, Tight Cramér–Rao type bounds for multiparameter quantum metrology through conic programming, Quantum 7, 1094 (2023).DOI
  20. [20]B. M. Terhal, Bell inequalities and the separability criterion, Phys. Lett. A 271, 319 (2000), arXiv:quant-ph/9911057.DOI
  21. [21]Y. C. Eldar and G. D. Forney, Jr., Optimal tight frames and quantum measurement, IEEE Trans. Inf. Theory 48, 599 (2002), arXiv:quant-ph/0106070.DOI
  22. [22]G. Chiribella and G. M. D’Ariano, Extremal covariant measurements, J. Math. Phys. 47, 092107 (2006), arXiv:quant-ph/0603168.DOI
  23. [23]J. Calsamiglia, M. Aspachs, R. Muñoz-Tapia, and E. Bagan, Phase-covariant quantum benchmarks, Phys. Rev. A 79, 050301(R) (2009), arXiv:0807.5126.DOI
  24. [24]D. S. Kaliuzhnyi-Verbovetskyi, I. M. Spitkovsky, and H. J. Woerdeman, Matrices with normal defect one, Oper. Matrices 3, 401 (2009), arXiv:0903.0090.
  25. [25]R. D. Wasson and H. J. Woerdeman, The normal defect of some classes of matrices, Linear Algebra Appl. 438, 3530 (2013), arXiv:1211.4145.DOI
  26. [26]K. Bierly, The normal Procrustes problem: A Riemannian optimization approach (2026), arXiv:2608.19513.arxiv.org/abs/2608.19513
  27. [27]N. Fabre, A. B. Klimov, G. Leuchs, and L. L. Sánchez-Soto, Local sampling of the SU(1,1) Wigner function, AVS Quantum Sci. 5, 014404 (2023).
  28. [28]Q. Wang, Nonorthogonal Nagaoka equality supports under Bayesian weak commutativity (2026), Zenodo preprint, doi:10.5281/zenodo.22083724.
  29. [29]C. K. K. Law and J. Z. Bernád, Bayesian approach to simultaneous quantum multiparameter estimation with finite data (2026), arXiv:2608.10230.arxiv.org/abs/2608.10230
  30. [30]D. Martínez, L. Pereira, K. Sawada, P. González, J. Cariñe, M. Muñoz, A. Delgado, E. S. Gómez, S. P. Walborn, and G. Lima, Experimental qudit state estimation via point tomography, Phys. Rev. Res. 7, 033119 (2025), arXiv:2412.14915.DOI
  31. [31]K. Glashoff and M. M. Bronstein, Matrix commutators: their asymptotic metric properties and relation to approximate joint diagonalization, Linear Algebra Appl. 439, 2503 (2013), arXiv:1305.2135.
  32. [32]V. Noferini and M. Zhukov, Closest normal matrix found again using Riemannian optimization (2026), arXiv:2608.28545.arxiv.org/abs/2608.28545
  33. [33]J. B. Lasserre, Global optimization with polynomials and the problem of moments, SIAM J. Optim. 11, 796 (2001).DOI

Paper details

vlnc:W_26e5w9bf
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

OpenAI Codex, Anthropic Claude Fable, Anthropic Claude Opus

View statement
Looking for related papers…
Contents