Pith. sign in

REVIEW 3 major objections 5 minor 48 references

Reading Qubits with Sequential Weak Measurements: Limits of Information Extraction

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that for generic sequential weak measurements, the mutual information between the initial qubit state and the measurement record saturates below one bit, so perfect readout is impossible however long the record is.

desk verdict Solid numerical evidence for a saturation phenomenon in sequential weak-measurement readout, but the analytic argument for a general bound rests on an explicit conjecture and a small-efficiency extrapolation; the abstract oversells the rigor. read the letter →

arxiv 2512.14583 v2 pith:5V3WUJUV submitted 2025-12-16 quant-ph cond-mat.stat-mech

Cesar Lema , Aleix Bou-Comas , Atithi Acharya , Vadim Oganesyan , Anirvan Sengupta This is my paper · ORCID
classification quant-phcond-mat.stat-mech MSC 81P1581P4794A17 PACS 03.65.Ta03.67.-a
keywords weakmeasurementqubitreadoutmutualinformationquantumtrajectoriesextractionrecordBayesoptimalestimationlimits
topics The Measurement Problem
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish a fundamental limit on reading out a qubit's initial state from a time-ordered record of weak measurements: for generic measurement schemes, the mutual information between the initial state and the measurement record saturates at a value strictly below one bit, so perfect recovery is impossible no matter how long the record is. It argues that this saturation is a generic feature of non-commuting weak measurement plus intrinsic dynamics, visible as a plateau in the scaling function f(x²T, φ/x²). The plateau value can be estimated analytically in a low-efficiency expansion, and it matches numerics. This matters because it sets an upper bound on any readout procedure, including machine-learning-based ones, and identifies an optimal measurement duration before extra data becomes noise.

What carries the argument

The averaged single-step measurement channel E — the superoperator obtained by averaging the Kraus operators over measurement outcomes — and its subleading eigenvalue, which defines an exponential correlation length ξ via e^{-1/ξ} = max_{λ≠1}|λ|. Finite ξ means late measurements carry exponentially little information about the initial state. A second piece is the continuum scaling limit (T∼x^{-2}, ϕ∼x²) that collapses the numerical data onto scaling curves, and a third is the low-efficiency (η≪1) approximation that reduces the readout problem to a binary-input additive white Gaussian noise channel whose signal-to-noise ratio γ(t) yields an explicit mutual-information plateau.

What would settle it

Compute exactly (e.g., by direct numerical simulation of the full channel) the mutual information I(S,A1:T) for Model II with ϕ≠0 at T ≫ ξ, with x small but finite; if it exceeds the plateau value and approaches log 2 — or any single non-commuting scheme with finite ξ achieves perfect asymptotic recovery — the saturation claim fails. Alternatively, an explicit construction of simultaneously diagonalizable but non-identity Kraus operators with an invariant subspace of dimension less than 2 that still yields mutual information reaching log 2 would refute the conjecture.

Watch

Extended reading notes

Core claim

The central claim is that, under generic circumstances, information about the initial qubit state effectively saturates past a certain number of observations and does not allow perfect recovery. Concretely, in the scaling limit the mutual information obeys I(S,A1:T|x,φ) ≈ f(x²T, φ/x²), and for generic non-commuting cases lim_{b→∞} f(b,a) < log 2. As a result, the Bayes-optimal readout fidelity is strictly below 100% even with perfect knowledge of the dynamics and an unlimited record. The paper also shows that late measurements are nearly independent of the initial state, decaying exponentially with a correlation length ξ, and conjectures that finite ξ implies imperfect recovery. An analytic

Load-bearing premise

The conjecture, stated after Eq. (24), that a finite correlation length ξ always implies imperfect recovery of initial-state information from an arbitrarily long measurement sequence.

Editorial extensions

If this is right

  • Beyond a timescale set by ξ, additional weak measurements add negligible information about the initial state; there is an optimal measurement duration for readout.
  • Any readout scheme, including machine learning classifiers, has accuracy strictly below 100% for generic non-commuting weak-measurement schemes, by Fano's inequality from the mutual-information bound.
  • Physics-agnostic supervised learning on long records overfits to late measurements that are effectively independent of the initial state; a Bayes-optimal classifier aware of the dynamics avoids this.
  • In some regimes weaker measurements extract more information than projective ones, producing nonmonotonic behavior in measurement strength.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The saturation bound likely extends to multi-qubit readout: any fixed non-commuting weak-measurement record defines a finite correlation length, so per-qubit information is bounded below the Holevo limit; the efficient-sampling mutual-information estimator in the paper is already adapted to that setting.
  • The correlation length ξ, computed from the Lindblad-averaged channel, could serve as a practical design criterion: hardware tuned so that the subleading eigenvalue of E approaches 1 should push the plateau toward one bit, while strongly non-commuting dynamics lowers it.
  • The plateau phenomenon is plausibly the measurement-side counterpart of dynamical purification and learnability transitions in monitored quantum circuits; a testable extension would be to look for a sharp learnability transition as the efficiency parameter η is tuned.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies how much information about a qubit's initial state can be extracted from a record of sequential weak measurements, focusing on two models: (I) informationally complete measurements of all three Pauli operators with no intrinsic dynamics, and (II) informationally incomplete Z-measurements with transverse-field unitary evolution. Using discrete-time simulations and continuous-time SME descriptions, the authors compute mutual information I(S,A_{1:T}) with Hoeffding-based error bounds, observe scaling collapse as a function of x^2 T (with φ/x^2 fixed), and identify plateaus below the ideal 1 bit. They supplement the numerics with analytic results: an exponential decay bound for the marginal mutual information of the T-th measurement characterized by a correlation length ξ, and a low-efficiency (η≪1) perturbative calculation of the mutual information plateau using a binary-input AWGN channel approximation. They connect these results to the Bayes-optimal readout and to overfitting in physics-agnostic learning. The central claim is that, generically, the measurement record does not allow perfect recovery of the initial state even as T→∞.

Significance. If the central claim holds, the paper establishes an important information-theoretic limitation for weak-measurement-based qubit readout that goes beyond specific estimation schemes. The numerical work is careful: MI estimates carry explicit Hoeffding concentration bounds, the scaling collapse is demonstrated across multiple parameter values, and the model-II analysis addresses a realistic non-QND readout setting. The analytic γ(t) formulas in Sec. IV C are parameter-free predictions (given τ, η, ω, φ) and their agreement with simulations in Fig. 3 is a genuine strength. The connection between information saturation and overfitting in supervised readout is practically relevant. The main weakness is that the analytic route to the saturation claim is incomplete: the finite-ξ conjecture in Sec. IV B is explicitly unproved, and the low-efficiency expansion is extrapolated to η=1 without control. Thus the paper is a valuable contribution with a credible central phenomenon, but the claimed 'bounds' are not yet fully established.

major comments (3)
  1. [Sec. IV B, after Eq. (24)] The central saturation claim—that finite ξ implies imperfect recovery—is stated as a conjecture: 'we conjecture that, in our problem, ξ being finite implies imperfect recovery of information about S from arbitrarily long sequence of measurements.' The preceding calculation bounds only the marginal mutual information I(S,A_T)=O(T e^{-2T/ξ}), not the total I(S,A_{1:T}). By the chain rule, Eq. (21), the total MI includes conditional terms I(S,A_j|A_{j+1:T}) that are not controlled by the marginal decay. Therefore no upper bound on the total MI follows from finite ξ. The abstract's phrase 'bounds on information extraction' overstates what is derived. To support the claim, the authors need either a proof of the conjecture (e.g., via a data-processing argument for the conditional terms) or a clearly stated weakening of the claim to 'numerical evidence suggests' for the specific models.
  2. [Sec. IV C and Fig. 3] The analytic mutual-information plateaus are obtained from a perturbative expansion in √η about η=0 (Eq. (29) and Appendix E) and then compared to simulations at η=1 in Fig. 3, with model II also at η=0.89, 0.5, 0.1. This is an uncontrolled extrapolation: the expansion parameter √η is not small at η=1. The agreement in Fig. 3(a,e) is empirical evidence, not a derivation. Calling these results 'bounds' (as in the abstract) is not justified; they are approximations. The manuscript should explicitly state this limitation and avoid the word 'bounds' unless a rigorous inequality is proven.
  3. [Sec. III B and Sec. IV A] The general claim that 'under generic circumstances' information saturates is inferred from two specific models, one with no intrinsic dynamics (Model I) and one with a single-axis measurement plus rotation (Model II). The conjecture in Sec. IV A about non-commuting Kraus operators is plausible but not established. The scaling function f(x^2T, φ/x^2) and the limit lim_{b→∞}f(b,a) are partly based on the 'naive scaling' assumption stated in Sec. III B. While the numerical collapse is convincing for the studied parameter range, the extrapolation to infinite T is not proven. The paper should carefully delimit the generality of the claim to the models and parameter ranges studied, or provide additional argument for the generic-case statement.
minor comments (5)
  1. [Eq. (24) and Appendix B] The definition of ξ via e^{-1/ξ}=max_{λ≠1}|λ| appears in Eq. (24) and again in Appendix B Eq. (B6). The notation is consistent, but the main text might benefit from a pointer to the derivation of the O(T e^{-2T/ξ}) bound, which is only sketched in Sec. IV B and detailed in Appendix B 1.
  2. [Appendix D, Eq. (D1)] The error-kernel noise model p_success=(1+(n−1)√η)/n is a specific choice. The paper should note that other noise models could lead to different efficiency parameterizations; this would help the reader judge the generality of the η-dependence.
  3. [Sec. IV C, Eq. (32)] In Eq. (32), the logarithms are presumably base 2 to give bits, but it is not explicitly stated. Please clarify the base of log in Eq. (32) and in the surrounding text.
  4. [Throughout] Several typos: 'eignevalues' in Appendix B; 'the curved in (e-h) peal off the plateau' in Sec. III B; 'GSKL' should be 'GKSL' (Gorini–Kossakowski–Sudarshan–Lindblad). Also, in Sec. I B, 'information complete' should be 'informationally complete' for consistency.
  5. [Fig. 3] The ratio plots in Fig. 3(a,e) would be more informative with quantitative error bars or shaded regions reflecting the Hoeffding bounds of the numerical estimates, since the claim of agreement is based on these ratios.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the MI estimates are parameter-free predictions checked against independent numerics; acknowledged conjectures are rigor gaps, not circular reductions.

full rationale

The paper's main analytic route is the small-efficiency perturbative SME expansion (Sec. IV C / Appendix E): rho_t is expanded in sqrt(eta), the zeroth-order dynamics is the Lindblad evolution, and the first-order measurement output is treated as a binary-AWGN signal whose SNR gamma(t) is computed from the difference of the two Lindblad solutions. These gamma(t) formulas (Eqs. 34-37) contain no free parameters fitted to the target mutual information; given tau, eta, omega, phi they are fixed predictions, and Fig. 3 checks them against independent discrete-time numerical MI estimates. Thus there is no fitted-input-called-prediction or definitional circle. There are no self-citations: references [8], [9], [14], etc. are external prior work, and the universal-weak-measurement Kraus operators are adopted as model definitions rather than used to force the saturation claim. The paper explicitly flags its main limitation: after Eq. (24) it states 'we conjecture that, in our problem, xi being finite implies imperfect recovery of information about S from arbitrarily long sequence of measurements,' and at the start of Sec. IV it says 'we do not have a rigorous way of establishing the conditions under which such information loss in measurement records occurs.' A labeled conjecture is an incompleteness in the derivation, not circularity. Likewise, evaluating the eta<<1 expansion at eta=1 is an uncontrolled extrapolation and overstates the word 'bounds,' but that is a validity/rigor concern, not a circular reduction. The central numerical plateau findings remain independent evidence.

Assumptions & free parameters 0 free parameters · 7 assumptions · 0 invented entities

No numerical constants are fitted to the target MI: the parameters x, τ, ω, φ, and η are model inputs and are varied. The analytic plateau formula is a parameter-free prediction of the SME/Lindblad dynamics once these inputs are fixed. The main burdens are the modeling assumptions listed above, especially the scaling assumption and the finite-ξ conjecture.

assumptions (7)
  • standard math Kraus operator/POVM representation and the validity of the discrete-time quantum trajectory description.
    Used in Section II A to define measurement-record probabilities and post-measurement states.
  • domain assumption Diffusive SME in the weak-measurement limit (x≪1) with the specific scaling T~x^{-2}, φ~x^2.
    Section II C and Appendix C; the continuum limit assumes these scalings so that a fixed SME with timescale τ and Hamiltonian coefficient ω exists.
  • ad hoc to paper The naive scaling assumption that MI depends on x and T only through x²T and on φ only through φ/x².
    Section III B; stated as 'reasonable to assume' and load-bearing for the plateau/scaling-collapse analysis.
  • ad hoc to paper The error-kernel noise model for inefficient measurements with p_success = (1+(n−1)√η)/n.
    Appendix D; defines the efficiency parameter η used in the perturbative calculation.
  • ad hoc to paper The conjecture that finite correlation length ξ implies imperfect recovery of the initial state from the full record.
    Section IV B; explicitly stated as a conjecture and used to support the general claim of information loss.
  • domain assumption Prior restricted to the two orthogonal Z-basis states |↑⟩, |↓⟩ with equal probability.
    Section II E and III; the claimed limits are for these optimal two-state priors, not for general initial states.
  • ad hoc to paper The low-efficiency perturbative expansion in √η and the Gaussian treatment of the output channel (bi-AWGN).
    Appendix E; an approximation whose leading term is used to compute the MI plateau and later extrapolated to η=1.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Reading Qubits with Sequential Weak Measurements: Limits of Information Extraction." pith.science (2026). https://pith.science/paper/5V3WUJUV

@misc{pith2026251214583,
  author       = {Pith},
  title        = {Pith review of: Reading Qubits with Sequential Weak Measurements: Limits of Information Extraction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/5V3WUJUV}},
  note         = {Machine review of arXiv:2512.14583}
}
read the original abstract

Quantum information processing and computation requires high accuracy qubit configuration readout. In many practical schemes, the initial qubit configuration has to be inferred from readout that is a time-dependent weak measurement record. However, a combination of the measurement scheme and intrinsic dynamics can end up scrambling the initial state and lose information irretrievably. Here, we study the information physics of quantum trajectories based on weak measurements in order to address the optimal achievable performance in qubit configuration readout for two realistic models of single qubit readout: (i) Model I is informationally complete, but without intrinsic dynamics; (ii) Model II is informationally incomplete weak measurements with intrinsic dynamics. We first use mutual information to characterize how much intrinsic information about the initial state is encoded in the measurement record. Using a fixed discrete time-step formulation, we compute the mutual information while varying the measurement strength, duration of measurement record, and the relative strength of intrinsic dynamics in our measurement schemes. We also exploit the emergence of continuum scaling and the Stochastic Master Equation in the weak measurement limit. We develop an asymptotic expansion in the measurement efficiency parameter to calculate mutual information, which captures qualitative and quantitative features of the numerical data. The bounds on information extraction are manifested as plateaux in mutual information, our analysis obtains these bounds and also optimal duration of measurement required to saturate them. Our results should be useful both for quantum device operation and optimization and also, possibly, for improving the performance of recent machine learning approaches for qubit and multiqubit configuration readout in current Noisy Intermediate-Scale Quantum (NISQ) experiment regimes.

Figures

Figures reproduced from arXiv: 2512.14583 by the authors.

Figure 1
Figure 1. FIG. 1: Numerical estimates of mutual information [PITH_FULL_IMAGE:figures/full_fig_p006_1.png] view at source ↗
Figure 2
Figure 2. FIG. 2: Nonmonotonic dependence of mutual [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. FIG. 3: Comparisons of numerical simulations and analytic (Sec. IV C) calculations of mutual information MI [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: FIG. 4: Prediction accuracy of Bayes optimal predictor [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 1 linked inside Pith

  1. [1]

    In this case, we weakly measure all Pauli operatorsσ∈ {X, Y, Z}

    Model I: Informationally complete measurements First, we introduce an informationally complete weak measurement model aiming to learn everything about an initial qubit state without any unitary dynamics. In this case, we weakly measure all Pauli operatorsσ∈ {X, Y, Z}. This measurement model is described by the set of six Kraus operatorsK σ y (x), which ar...

  2. [2]

    We con- sider an informationally incomplete weak measurement, equivalent to measuring a single observable and the uni- versal weak measurement in Eq

    Model II: Informationally incomplete measurements with unitary evolution Our second model, based on the work of [9], corre- sponds to the generic case of measuring an observable with non-commuting Hamiltonian evolution. We con- sider an informationally incomplete weak measurement, equivalent to measuring a single observable and the uni- versal weak measur...

  3. [3]

    The saturation of mutual information with increas- ingT,

  4. [4]

    We will restrict our analysis to the qubit setting

    The emergence of scaling with smallxand largeT, consistent with the continuum limit. We will restrict our analysis to the qubit setting. For initial states, we use the (pure) up and down states (the Pauli-Zbasis states), thus referring toP 0 asS. These states represent the optimal setting for information ex- traction: two orthogonal states, for which the ...

  5. [5]

    7 (a) x (b) x √ T FIG

    continuous time low efficiency limit, we develop a per- turbative approach to estimate the mutual information plateau of the weak measurement limit in general set- tings. 7 (a) x (b) x √ T FIG. 2: Nonmonotonic dependence of mutual information (MI)I(S;A 1:T ) on measurement (x) and precession field (ϕ) strengths outside of the scaling limit in Model II. Pa...

  6. [6]

    Krantz, M

    P. Krantz, M. Kjaergaard, F. Yan, T. P. Orlando, S. Gus- tavsson, and W. D. Oliver, A quantum engineer’s guide to superconducting qubits, Applied Physics Reviews6, 021318 (2019)

    2019

  7. [7]

    Leonhardt,Measuring the quantum state of light, Vol

    U. Leonhardt,Measuring the quantum state of light, Vol. 22 (Cambridge university press, 1997)

    1997

  8. [8]

    Gambetta, W

    J. Gambetta, W. Braff, A. Wallraff, S. Girvin, and R. Schoelkopf, Protocols for optimal readout of qubits us- ing a continuous quantum nondemolition measurement, Physical Review A76, 012325 (2007)

    2007

Show all 48 references
  1. [9]

    Angelatos, S

    G. Angelatos, S. A. Khan, and H. E. T¨ ureci, Reser- voir computing approach to quantum state measurement, Physical Review X11, 041062 (2021)

    2021

  2. [10]

    R. B. Ash,Information theory(Courier Corporation, 2012)

    2012

  3. [11]

    T. M. Cover and J. A. Thomas,Elements of Information Theory 2nd Edition (Wiley Series in Telecommunications and Signal Processing)(Wiley-Interscience, 2006)

    2006

  4. [12]

    Meyen,Relation between classification accuracy and mutual information in equally weighted classification tasks, Ph.D

    S. Meyen,Relation between classification accuracy and mutual information in equally weighted classification tasks, Ph.D. thesis, Universit¨ at Hamburg (2016)

    2016

  5. [13]

    Oreshkov and T

    O. Oreshkov and T. A. Brun, Weak measurements are universal, Physical review letters95, 110409 (2005)

    2005

  6. [14]

    Tang and F

    X. Tang and F. Li, Phase diagram of the dynamics of a precessing qubit under a quantum measurement, Physical Review A101, 062102 (2020)

    2020

  7. [15]

    D. I. Schuster,Circuit quantum electrodynamics(Yale University, 2007)

    2007

  8. [16]

    Jeffrey, D

    E. Jeffrey, D. Sank, J. Mutus, T. White, J. Kelly, R. Barends, Y. Chen, Z. Chen, B. Chiaro, A. Dunsworth, et al., Fast accurate state measurement with super- conducting qubits, Physical review letters112, 190504 (2014)

    2014

  9. [17]

    Barchielli and M

    A. Barchielli and M. Gregoratti,Quantum trajectories and measurements in continuous time: the diffusive case, Vol. 782 (Springer, 2009)

    2009

  10. [18]

    Rouchon, A tutorial introduction to quantum stochas- tic master equations based on the qubit/photon system, Annual Reviews in Control54, 252 (2022)

    P. Rouchon, A tutorial introduction to quantum stochas- tic master equations based on the qubit/photon system, Annual Reviews in Control54, 252 (2022)

    2022

  11. [19]

    C. Lin, Y. Ma, and D. Sels, Asymptotic behav- ior of continuous weak measurement and its applica- tion to real-time parameter estimation, arXiv preprint arXiv:2311.02066 (2023)

    2023 arXiv

  12. [20]

    Gorini, A

    V. Gorini, A. Kossakowski, and E. C. G. Sudarshan, Completely positive dynamical semigroups of n-level sys- tems, Journal of Mathematical Physics17, 821 (1976)

    1976

  13. [21]

    Lindblad, On the generators of quantum dynamical semigroups, Communications in mathematical physics 48, 119 (1976)

    G. Lindblad, On the generators of quantum dynamical semigroups, Communications in mathematical physics 48, 119 (1976)

    1976

  14. [22]

    Hoeffding, Probability inequalities for sums of bounded random variables, The collected works of Wass- ily Hoeffding , 409 (1994)

    W. Hoeffding, Probability inequalities for sums of bounded random variables, The collected works of Wass- ily Hoeffding , 409 (1994)

    1994

  15. [23]

    K. P. Murphy,Probabilistic machine learning: an intro- duction(MIT press, 2022)

    2022

  16. [24]

    M. J. Gullans and D. A. Huse, Dynamical purifica- tion phase transition induced by quantum measurements, Physical Review X10, 041020 (2020)

    2020

  17. [25]

    Ippoliti and W

    M. Ippoliti and W. W. Ho, Dynamical purification and the emergence of quantum state designs from the pro- jected ensemble, PRX Quantum4, 030322 (2023)

    2023

  18. [26]

    Ippoliti and V

    M. Ippoliti and V. Khemani, Learnability transitions in monitored quantum dynamics via eavesdropper’s classi- cal shadows, PRX Quantum5, 020304 (2024)

    2024

  19. [27]

    Choi,Adaptive and iterative signal processing in com- munications(Cambridge University Press, 2006)

    J. Choi,Adaptive and iterative signal processing in com- munications(Cambridge University Press, 2006). 12 Appendix A: Estimating Mutual Information and Initial State Estimation Accuracy In our analysis, we used mutual information as the central measure of the intrinsic inform...

    2006

  20. [28]

    Mutual information estimation We can accurately estimate mutual informationI(S, A 1:T ) between the initial bitSand the measurement record A1:T using Hoeffding’s inequality. First, consider the mutual information as, I(S, A1:T ) =H(S)−H(S|A 1:T ) =H(S)− X a1:T Pr [A1:T =a 1:T ...

  21. [29]

    In initial state estimation, our objective is to classify the initial state correctly given a measurement recorda 1:T

    Estimating qubit initial state estimation accuracy Now, we look at estimating the accuracy of our Bayes optimal predictor in single qubit initial state recov- ery/estimation. In initial state estimation, our objective is to classify the initial state correctly given a measurem...

  22. [30]

    Initial state andT-th measurement joint probability In the both accuracy and in mutual information, we see that in the generic case, our ability to infer the initial state from the sequential measurements remains limited, even whenTgoes to infinity. To understand when very lat...

  23. [31]

    Informationally complete measurements Now we calculateξfor our first measurement scheme. First we can calculate the individual measurement matrices as, EX,y = 1 6   1ytanhx0 0 ytanhx1 0 0 0 0 sechx0 0 0 0 sechx   , EY,y = 1 6   1 0ytanhx0 0 sechx0 0 ytanhx0 1 0 0 0 0 ...

  24. [32]

    Informationally incomplete measurements with unitary evolution For our next measurement model, as usual, defineϕ:=ω∆t. We calculate the individual measurement matricies as, Ey =   1ytanhxsinϕ0ytanhxcosϕ 0 sechxcosϕ0−sechxsinϕ 0 0 sechx0 ytanhxsinϕ0 cosϕ   (B9) and, after...

  25. [33]

    Stochastic master equation in the weak measurement limit In general, the continuous time dynamics of our weak measurement schemes are Stochastic Master Equations in the diffusive limit [12, 13], given by, dρt =−i h H0, ρt i dt+ 1 τ X ν LνρtL† ν − 1 2 L† νLνρt +ρ tL† νLν ! dt +...

  26. [34]

    We specifically consider measurements of the single-qubit Pauli operatorsσ∈ {X, Y, Z}

    Sequential weak measurements Our sequential weak measurement schemes are derived from universal weak measurements [8] of observablesOwith eigenvalues in the set{1,−1}. We specifically consider measurements of the single-qubit Pauli operatorsσ∈ {X, Y, Z}. In the sequential meas...

  27. [35]

    NX k=1 ϵyk ρ # = 2N ϵ2Tr [σρ] (C15) E   NX k=1 ϵyk !2 ρ   =E

    Informationally incomplete weak measurements We first derive the Stochastic Master Equation (SME) in the context of an informationally incomplete weak mea- surement scheme, equivalent to universal weak measurements. Our construction is the sequence of weak Pauli mea- surements...

  28. [36]

    In this case, we model the measurement outcomes as vectors⃗ y∈R3, where each componenty σ ∈ {+1,0,−1}corresponds to a measurement outcome along theσaxis

    Informationally Complete Measurements Next we derive the SME in the informationally complete measurement scheme where we consider simultaneous weak measurements of the qubit PauliX,Y, andZoperators. In this case, we model the measurement outcomes as vectors⃗ y∈R3, where each c...

  29. [37]

    Informationally Incomplete W eak Measurements with Unitary Dynamics Finally, we derive the SME with unitary dynamics. Again, consider weak measurements of the qubit Pauli-Zoper- ator, the same as the informationally incomplete measurements in Appendix section C 3, but now unde...

  30. [38]

    Expansion in the efficiency parameter The continuous time dynamics of a quantum state in the weak measurement limit with noisy measurements is given generally by the SME C1 and measurement output C4. Without loss of generality, consider this general SME, including an explicit ...

  31. [39]

    Then, we can use the solutions to the approximate dynamics (Eqs

    Mutual information in the small efficiency limit In the small efficiency limit, √η≪0, we can approximate the measurement SME dynamics in small orders of efficiency. Then, we can use the solutions to the approximate dynamics (Eqs. E5 and E6) to estimate the mutual information b...

  32. [40]

    It is defined with the (noisy observations) outputY t = √ρXt +Z t with (clean signal) inputX t ∈ {±1} and (noise)Z t ∼ N(0,1) withZt being independent for eacht

    Mutual information for the bi-A WGN channel The binary Additive Gaussian White Noise (bi-A WGN) is a classical communication channel of noisy observations of a clean signal. It is defined with the (noisy observations) outputY t = √ρXt +Z t with (clean signal) inputX t ∈ {±1} a...

  33. [41]

    25, by solving for the Lindblad dynamics, in the small efficiency limit, and using the bi-A WGN channel

    Informationally complete measurements Here, we estimate the mutual information in our informationally complete weak measurement model, Eq. 25, by solving for the Lindblad dynamics, in the small efficiency limit, and using the bi-A WGN channel. The Lindblad Master equation for ...

  34. [42]

    Informationally incomplete measurements with unitary dynamics For our next model, informationally incomplete measurements with unitary dynamics, Eq. 27, the Lindblad Master equation is, dρ(0) t =−i ω 2 [X, ρ(0) t ]dt+ 1 τ Zρ (0) t Z−ρ (0) t dt(E20) We again solve this in the P...

  35. [43]

    Forα > 1 2 , letµ := √ 4α2 −1, ρy(t) = e− 1 τ t µ h µcos µ τ t −sin µ τ t ρy(0)−2αsin µ τ t ρz(0) i (E26) ρz(t) = e− 1 τ t µ h µcos µ τ t + sin µ τ t ρz(0) + 2αsin µ τ t ρy(0) i (E27)

  36. [44]

    Forα < 1 2 , letµ := √ 1−4α 2, ρy(t) = e− 1 τ t µ h µcosh µ τ t −sinh µ τ t ρy(0)−2αsinh µ τ t ρz(0) i (E28) ρz(t) = e− 1 τ t µ h µcosh µ τ t + sinh µ τ t ρz(0) + 2αsinh µ τ t ρy(0) i (E29)

  37. [45]

    Forα= 1 2 , ρy(t) =e − 1 τ t 1− t τ ρy(0)− t τ ρz(0) (E30) ρz(t) =e − 1 τ t t τ + 1 ρz(0) + t τ ρy(0) (E31) Solving for initial up and down spin states ρ(0)↑(0) = 1 2 [I+Z] andρ (0)↓(0) = 1 2 [I−Z] we get,

  38. [46]

    I+ −2αe− 1 τ t µ sin µ τ t Y+ e− 1 τ t µ µcos µ τ t + sin µ τ t Z # ρ(0)↓(t) = 1 2

    Forα > 1 2 : ρ(0)↑(t) = 1 2 " I+ −2αe− 1 τ t µ sin µ τ t Y+ e− 1 τ t µ µcos µ τ t + sin µ τ t Z # ρ(0)↓(t) = 1 2 " I+ 2αe− 1 τ t µ sin µ τ t Y+ −e− 1 τ t µ µcos µ τ t + sin µ τ t Z # 25 and the difference is, ρ(0)↑(t)−ρ (0)↓(t) = " −2αe− 1 τ t µ sin µ τ t Y+ e− 1 τ t µ µcos µ ...

  39. [47]

    I+ −2αe− 1 τ t µ sinh µ τ t Y+ e− 1 τ t µ µcosh µ τ t + sinh µ τ t Z # (E34) ρ(0)↓(t) = 1 2

    Forα < 1 2 : ρ(0)↑(t) = 1 2 " I+ −2αe− 1 τ t µ sinh µ τ t Y+ e− 1 τ t µ µcosh µ τ t + sinh µ τ t Z # (E34) ρ(0)↓(t) = 1 2 " I+ 2αe− 1 τ t µ sinh µ τ t Y+ −e− 1 τ t µ µcosh µ τ t + sinh µ τ t Z # (E35) and the difference is, ρ(0)↑(t)−ρ (0)↓(t) = " −2αe− 1 τ t µ sinh µ τ t Y+ e−...

  40. [48]

    E9 for theseγ(t)

    Forα= 1 2 : ρ(0)↑(t) = 1 2 I+ − t τ e− 1 τ t Y+ 1 + t τ e− 1 τ t Z (E37) ρ(0)↓(t) = 1 2 I+ t τ e− 1 τ t Y+ − 1 + t τ e− 1 τ t Z (E38) and the difference is, ρ(0)↑(t)−ρ (0)↓(t) = − t τ e− 1 τ t Y+ 1 + t τ e− 1 τ t Z (E39) Then the SNRγ(t) is, γ(t) =η 5−e − 2 τ t 2 t τ 2 + 6 t τ...

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.