We quantify the extra Bayes risk of locating a quantum change point and estimating its amplitude with one measurement. The joint and marginal optimizations use the full record and the same joint prior, with both parameters unknown. For product records on a known pure state curve, position has clipped absolute loss in resolution units, and amplitude offset has squared loss in fixed calibrated units. Ideal steps exhibit certified conflict at finite signal strength, including in phase records of sixteen qubits, and recover both informative marginal optima at strong signal. For finite offset prior width, a retained quantum boundary region serves position, and the remaining suffix serves amplitude. Physical recovery holds when strength grows more slowly than the square root of record length. Known smooth rises also admit leading recovery in the coherent experiment as strength grows and width shrinks. For the quintic profile, we establish two residual cost scaling laws. The residual cost is $Θ(s^{-2})$ as $s\to\infty$. Taking the strong signal limit at fixed $w$ yields a limiting cost of $Θ(w^3)$ as $w\to0$. In the resolved local model, quadratic position loss admits a common optimum within the covariant class. These results link onset geometry and decision loss to Bayesian measurement compatibility in quantum change point estimation.
AI Summary
Can one measurement optimally locate a quantum signal’s onset and estimate its amplitude when both are unknown? This paper studies product records along a known pure-state curve, with uniform position and amplitude-offset priors. It defines a compatibility cost as the weighted joint Bayes optimum minus the weighted marginal optima, all using the full record and same joint prior. Position uses resolution-scaled clipped absolute loss; amplitude offset uses squared loss.
For ideal steps, certified optimization gives $1.00\times10^{-3}<\mathcal G_{1/2}<1.84\times10^{-3}$ at a reference strength in the coherent comparison model. A 16-qubit phase-record certificate establishes positive cost across a finite prior-width interval, amounting to $0.58\%$–$2.76\%$ of the weighted marginal baseline. Conversely, both marginal excess risks vanish at strong signal: coarse localization retains a quantum boundary region for position and allocates the remaining suffix to amplitude. Physical recovery holds for $s_n\to\infty$ with $s_n=o(\sqrt n)$, without making either task uninformative.
Known smooth rises also recover compatibility as strength grows and width shrinks. For a quintic profile and specified prior, the residual cost is $\Theta(s^{-2})$ uniformly for $s^2w\in[1/2,2]$; taking strong signal first gives $\Theta(w^3)$. Proofs combine Bayesian operator criteria, coherent-experiment comparisons, decision-theoretic bounds, and numerical certificates. Quadratic position loss instead permits compatibility within the resolved local covariant class. The results assume known onset shape and width; their uncertainty, broader residual-scaling universality, and finite-record implementation and memory requirements remain open.
Little sign of AI-generated textStrong assumptions
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
We quantify the extra Bayes risk of locating a quantum change point and estimating its amplitude with one measurement. The joint and marginal optimizations use the full record and the same joint prior, with both parameters unknown. For product records on a known pure state curve, position has clipped absolute loss in resolution units, and amplitude offset has squared loss in fixed calibrated units. Ideal steps exhibit certified conflict at finite signal strength, including in phase records of sixteen qubits, and recover both informative marginal optima at strong signal. For finite offset prior width, a retained quantum boundary region serves position, and the remaining suffix serves amplitude. Physical recovery holds when strength grows more slowly than the square root of record length. Known smooth rises also admit leading recovery in the coherent experiment as strength grows and width shrinks. For the quintic profile, we establish two residual cost scaling laws. The residual cost is Θ(s−2) as s→∞. Taking the strong signal limit at fixed w yields a limiting cost of Θ(w3) as w→0. In the resolved local model, quadratic position loss admits a common optimum within the covariant class. These results link onset geometry and decision loss to Bayesian measurement compatibility in quantum change point estimation.
Can one measurement optimally locate a quantum signal’s onset and estimate its amplitude when both are unknown? This paper studies product records along a known pure-state curve, with uniform position and amplitude-offset priors. It defines a compatibility cost as the weighted joint Bayes optimum minus the weighted marginal optima, all using the full record and same joint prior. Position uses resolution-scaled clipped absolute loss; amplitude offset uses squared loss.
For ideal steps, certified optimization gives 1.00×10−3<G1/2<1.84×10−3 at a reference strength in the coherent comparison model. A 16-qubit phase-record certificate establishes positive cost across a finite prior-width interval, amounting to 0.58%–2.76% of the weighted marginal baseline. Conversely, both marginal excess risks vanish at strong signal: coarse localization retains a quantum boundary region for position and allocates the remaining suffix to amplitude. Physical recovery holds for sn→∞ with sn=o(n), without making either task uninformative.
Known smooth rises also recover compatibility as strength grows and width shrinks. For a quintic profile and specified prior, the residual cost is Θ(s−2) uniformly for s2w∈[1/2,2]; taking strong signal first gives Θ(w3). Proofs combine Bayesian operator criteria, coherent-experiment comparisons, decision-theoretic bounds, and numerical certificates. Quadratic position loss instead permits compatibility within the resolved local covariant class. The results assume known onset shape and width; their uncertainty, broader residual-scaling universality, and finite-record implementation and memory requirements remain open.
We determine how errors in a coherent receiver's outputs change the cost of calibrating an exponentially rare quantum component. For a smooth moment-matched phase family, one receiver exposes informative events that counting misses. The optimal robust cost follows an explicit curve: an ideal-cost plateau ends at a tail-mass threshold, then an increasing branch approaches the feasibility boundary. One stopping rule attains this exponent at every unknown label, even when conditional errors depend on the observed history. Near the boundary, the remaining contamination margin δ and an explicit midpoint variance v* determine a finer law. As δ and the error target α vanish, the minimax cost is 2v* δ⁻² log(1/α), without a relative-rate restriction for a fixed receiver. This coefficient governs fixed sampling and every fixed positive-moment cost of arbitrary randomized stopping; a single clean source forces the lower bound. A uniform theorem carries the constant to changing finite receivers under explicit margin conditions and gives a computable critical strip for the physical family. These laws locate the loss of efficient calibration and quantify the cost of approaching the last feasible regime. This author preprint includes the main manuscript, complete proofs, LaTeX sources and computational materials. Journal peer review has not been completed. The scientific files are preserved unchanged; author identity is provided in the record metadata and the accompanying README. Version V1.7. All files are embargoed until 1 February 2027, when they will automatically become publicly accessible. Bibliographic metadata are public.
Quantum error mitigation estimates expectation values that are often used to select or rank candidates. We study how estimation accuracy, decision risk and sampling cost are related in these tasks. The joint law of residual differences determines the risk of a shift-invariant decision; in Gaussian regimes, effective margins and their covariance provide a computable representation. This connects mitigation maps, shared measurement records and shot allocation, and yields an extrapolation design criterion under certified margins.
Finite-shot QAOA-MaxCut controls illustrate two forms of this relationship: a positive affine CDR correction reduces mean-squared error while preserving rankings, whereas signed PEC sampling reduces mean-squared error and increases selection failures at equal evaluation cost. A retrospective, oracle-assisted study of 48 landscapes compares Gaussian-risk and MSE-based method selection using separate fitting, selection and evaluation samples. At the primary total shot budget, with learning amortized over 20 decisions per landscape, failure rates are 0.3083 for Gaussian-risk selection, 0.4719 for MSE selection and 0.2333 for unmitigated estimation. These comparisons describe how uncertainty in candidate differences and the allocation of measurements between learning and evaluation jointly enter the assessment of mitigation.
AbstractRetained measurement records constrain recoverability, but their formation also depends on the available physical controls. We give exact, constructively attained answers to both questions in two specified models. For a balanced nondemolition qubit process with actual binary observer memories unavailable to the decoder, we determine the maximum recovery compatible with average readout accuracy and pair agreement. The finite-record solution includes its feasible region and a common memory-writing and recovery process, without assuming exchangeability of observed records. Under a complete conditional-pairwise-independence promise at fixed informative accuracy, the optimal coherence has a sharp inverse-square-root asymptotic; fully conditionally independent records instead give exponential decay. Thus identical single-record and pair-record calibrations need not justify the product-law recovery bound. We then fix a logical code in four spin-3/2 systems and sequential diagonal contacts with exact controlled-swap endpoints. Hard budgets for net contact action and added spectral control charge yield an exact peak-excursion frontier within a declared quadratic-plus-cubic control class. A single positive-amplitude preparation jointly attains the record, recovery and occupation requirements, connecting the two results through an exact finite-dimensional optimization. A closed three-record example distinguishes fixed complete records from fixed averaged data. Separately calibrated imperfect contacts admit one-sided finite-error extensions. Together, the results characterize both informational compatibility and control-dependent formation through common attaining apparatuses.Preprint status and related materialsAuthor-submitted preprint; not peer reviewed or accepted for publication. This theoretical and computational work addresses the compatibility of retained binary quantum records with optimized accessible-system recovery, and the control-dependent resources required to form those records in a specified spin encoding. The article provides exact finite-record bounds, attainable recovery and resource frontiers, and calibrated finite-error qualifications. The manuscript has been submitted to Foundations of Physics for journal consideration; this Figshare record is not a journal publication. The technical supplement is included as a second PDF in this same record. Related software, verification code, exact certificates, and selected computational evidence are available separately at https://doi.org/10.6084/m9.figshare.34886079 under the MIT licence. When a journal version of record becomes available, a link to its DOI will be added here.AuthorKent Nimmo; Independent Researcher, Vancouver, British Columbia, Canada; ORCID 0009-0003-8503-8891.Reading and related workBrowse Quantum Foundations for related work on records, access and measurement, and Mathematics for exact bounds and certified constructions. The MIT software/certificate archive remains a distinct record from these manuscript PDFs.
Recovery from a coherent rotation with an unobserved angle depends on both the encoding and what calibration reveals about the noise. We determine their combined limit for a known finite-dimensional generator and one final recovery. Two polynomial projections organize the answer. A physical Krylov chain separates correctable error amplitudes from their first state-dependent distortion; projection onto lower-angle powers identifies the part that a coherent decoder can cancel. Under explicit rank and moment conditions, their combination gives the exact leading optimal entanglement infidelity at weak noise and a recovery attaining it for any finite logical dimension. The optimal worst-input error lies between one and d times that infidelity at leading order, for logical dimension d. When the angle distribution is unknown, we optimize one decoder for every distribution compatible with the calibration. When spectral and calibration frequencies are integer multiples of a common frequency, moment positivity turns this into an exact finite optimization. Small measured means can still allow rare finite rotations. If the sole calibration datum is βr=E[sin2r(θ/2)], with θ the wrapped rotation phase at that common frequency, the optimal worst-prior infidelity is O(βr) exactly when generator powers through degree r−1 are correctable. When the first uncorrectable power has degree m<r, the optimal loss scales as βrm/r, with a sharp coefficient fixed by a universal moment problem. In a loss-free four-photon memory, population measurements at two calibration durations retain the fully informed quartic error coefficient for a fixed symmetric narrow-noise family with positive variance and finite fourth moment; a single duration permits loss quadratic in the noise width.
Results concerning the construction of quantum Bayesian error regions as a means to certify the quality of parameter point estimators have been reported in recent years. This task remains numerically formidable in practice for large dimensions and so far, no analytical expressions of the region size and credibility (probability of any given true parameter residing in the region) are known, which form the two principal region properties to be reported alongside a point estimator obtained from collected data. We first establish analytical formulas for the size and credibility that are valid for a uniform prior distribution over parameters, sufficiently large data samples and general constrained convex parameter estimation settings. These formulas provide a means to an efficient asymptotic error certification for parameters of arbitrary dimensions. Next, we demonstrate the accuracies of these analytical formulas as compared to numerically computed region quantities with simulated examples in qubit and qutrit quantum-state tomography where computations of the latter are feasible.
How much detected thermal light is needed to identify several emitting sites? This working paper studies binary sources on a known grid of addressable source-image modes, using one fixed passive optical mask, photon counting, and a predetermined observation length. It distinguishes the original one-reference apparatus, complete Fourier reference collection, and a separate apparatus with one added vacuum input. All recorded photons count toward the measurement cost. The central result establishes the optimal quadratic detected-photon order at unit brightness for every fixed support limit on every grid, without geometric restrictions or primality assumptions. A reciprocal transmission mask attains this order using target counts alone, while recorded reference photons remain included in the cost. The proof reconstructs short cyclic arcs from two rational-probe values, then uses contour integration to isolate entire clusters of poles. This gives uniform quantitative separation without requiring a fixed minimum distance between individual poles. A bounded candidate list removes the grid-size logarithm from the final comparisons. On prime grids, a shorter finite-field argument recovers the support from two integer edge profiles. The quadratic law also holds uniformly over every nonempty reference subset when known brightness remains in a fixed interval bounded away from zero. The contour argument extends to a general stability theorem for rational interaction observables: under explicit locality and conditioning assumptions, quantitative distinguishability within clusters yields quantitative separation of the global polynomial data. For irregular registered dictionaries, this gives a separation margin proportional to the minimum site spacing, with an optimal spacing exponent. This extension assumes exactly resolved distance channels; their optical acquisition and robustness to unknown position errors require further analysis. Explicit dependence on the support limit also yields a growing-support regime: the photon cost remains n2+o(1)log(1/δ) whenever s2log(2s)=o(logn). The accompanying computational bound is stated in real arithmetic, and the dependence on sparsity is deliberately conservative. The manuscript retains explicit ramp-mask constructions for every configuration of at most four emitters, for short-span sources, and for sources with distinct cyclic pair distances. The last class has no span restriction and contains an asymptotically overwhelming fraction of fixed-size supports. For four emitters, ordinary-correlation rigidity and cyclic separation together distinguish reflected configurations. An explicit passive mask also attains the exact optimal Fourier-intensity separation under a prescribed transmission floor. In both the original and complete-reference experiments, sufficiently faint light produces a different law: at fixed sparsity, the optimal photon order grows as the grid size raised to twice the support limit. For at most two emitters, the paper determines the optimal detected-photon order throughout the known-brightness range, from arbitrarily faint light through unit brightness, on odd grids in the original apparatus and on every grid with complete reference collection. The result includes the dependence on grid size, brightness, and error probability, including an additional logarithmic information cost between the faint and bright limits. At fixed confidence, the quadratic regime begins at brightness of order (logn)/n2. A mask combining an index ramp with a quadratic-residue code attains the sharp order. On even physical grids, one known vacuum input and odd-dimensional Fourier optics give the same law; omitting one target output preserves the detector count. This is a distinct apparatus. For the specified two-scale transmissions on an infinite even-grid family, an exact lower bound using the complete count record proves that changing phases or decoders alone cannot remove their intermediate-brightness limitation. A collected-light model makes the physical scope explicit. For a known collection operator, binary-source identifiability is equivalent to injectivity of the induced thermal covariance map. Covariance collisions give identical quantum states and therefore prevent discrimination by any subsequent receiver. A passive multiport establishes population identifiability when the map is injective. The paper also derives the exact throughput for uniform passive mode separation from the smallest Gram eigenvalue and accounts separately for detected and incident photons. These results connect the assumed mode access to a specified collection model; they do not establish a general quantum-optimal receiver or a complete astronomical implementation. Reference excess variance supplies a normalization that identifies the spatial source despite an unknown finite common-profile thermal-mode spectrum and independent unknown Poisson backgrounds. The finite-sampling guarantee includes the empty source under explicit mean and nonzero-signal bounds. Population identifiability of the mode spectrum is distinguished from stable estimation. When background means are known, the arm means recover source mass and improve a guaranteed scale-free precision radius. Further results include exact thermal count variances, confidence exponents, calibration bounds, a photon penalty for discarding detector labels, and an exact target-only localization obstruction when singleton brightness is unrestricted and unknown. The results have analytical proofs, supported by 34 reproducible finite and symbolic verification suites. These checks provide corroboration, not experimental validation or external peer review. The unrestricted fixed-sparsity unit-brightness problem is resolved on all grids. Remaining questions include the optimized even-grid intermediate-brightness transition in the original apparatus, transitions for three or more emitters, sharp growing-support rates, sharp reference-coverage dependence at vanishing brightness, and optimal covariance-only scale-free precision. Continuous off-grid recovery requires a tolerance-based criterion and additional stability analysis. The work makes no literature-priority claim.