Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Precision Limits of Hamiltonian Learning for Quantum Processors under Calibration Uncertainty

Yi Pan, Meng Hsiu Tsai, Hanqi Jiang, Junhao Chen, Wei Zhang, Isaac Lyngaas, Yingfeng Wang, Tianming Liu

Published Nov 12, 2026• 0 citations
PDF (CC BY)Publisher

Abstract

Abstract Learning the Hamiltonian of a quantum processor is important for controlling and certifying quantum devices, but the state preparation and measurement used to probe it are themselves imperfect. Even when exact data identify an interaction uniquely, identifiability alone does not determine finite-data precision. Here we show, for n-qubit Markovian dynamics under geometric regularity conditions, that the local uncertainty of a Hamiltonian coefficient scales either linearly or as the square root of the entrywise error in the generator matrix in the readout gauge. Square-root scaling occurs exactly when a gauge transformation moves the coefficient at first order while noise positivity can be restored at second order, and real noise correlations drive this first-order change. As a consequence, under suitable dynamical and readout conditions, square-root scaling forces uniformly valid confidence intervals from N shots to have expected length Ω(N⁻¹⁄⁴), even for adaptive local experiments. Under a common generator norm bound, estimating this matrix attains this rate up to logarithmic factors. Experiments on three superconducting processors show that additional measurement bases increase the visibility of injected rotations at a matched number of shots. Whether a coupling admits the usual N⁻¹⁄² scaling is therefore set by how noise correlations and positivity constrain its gauge directions.

DOI
10.21203/rs.3.rs-11695581/v1

Related papers

Scalable Bayesian Hamiltonian learning

T. J. Evans, Robin Harper, Steven T. Flammia

As the size of quantum devices continues to grow, the development of scalable methods to characterise and diagnose noise is becoming an increasingly important problem. Recent methods have shown how to efficiently estimate Hamiltonians in principle, but they are poorly conditioned and can only characterize the system up to a scalar factor, making them difficult to use in practice. In this work we present a Bayesian methodology, called Bayesian Hamiltonian Learning (BHL), that addresses both of these issues by making use of any or all, of the following: well-characterised experimental control of Hamiltonian couplings, the preparation of multiple states, and the availability of any prior information for the Hamiltonian. Importantly, BHL can be used online as an adaptive measurement protocol, updating estimates and their corresponding uncertainties as experimental data become available. In addition, we show that multiple input states and control fields enable BHL to reconstruct Hamiltonians that are neither generic nor spatially local. We demonstrate the scalability and accuracy of our method with numerical simulations on up to 100 qubits. These practical results are complemented by several theoretical contributions. We prove that a $k$-body Hamiltonian $H$ whose correlation matrix has a spectral gap $Δ$ can be estimated to precision $\varepsilon$ with only $\tilde{O}\bigl(n^{3k}/(\varepsilon Δ)^{3/2}\bigr)$ measurements. We use two subroutines that may be of independent interest: First, an algorithm to approximate a steady state of $H$ starting from an arbitrary input that converges factorially in the number of samples; and second, an algorithm to estimate the expectation values of $m$ Pauli operators with weight $\le k$ to precision $ε$ using only $O(ε^{-2} 3^k \log m)$ measurements, which quadratically improves a recent result by Cotler and Wilczek.

arXiv:1912.07636•Dec 16, 2019•113 citations

Robustly learning the Hamiltonian dynamics of a superconducting quantum processor

Dominik Hangleiter, Ingo Roth, Jonáš Fuksa +2 more

Abstract Precise means of characterizing analog quantum simulators are key to developing quantum simulators capable of beyond-classical computations. Here, we precisely estimate the free Hamiltonian parameters of a superconducting-qubit analog quantum simulator from measured time-series data on up to 14 qubits. To achieve this, we develop a scalable Hamiltonian learning algorithm that is robust against state-preparation and measurement (SPAM) errors and yields tomographic information about those SPAM errors. The key subroutines are a novel super-resolution technique for frequency extraction from matrix time-series, tensorESPRIT, and constrained manifold optimization. Our learning results verify the Hamiltonian dynamics on a Sycamore processor up to sub-MHz accuracy, and allow us to construct a spatial implementation error map for a grid of 27 qubits. Our results constitute an accurate implementation of a dynamical quantum simulation that is precisely characterized using a new diagnostic toolkit for understanding, calibrating, and improving analog quantum processors.

arXiv:2108.08319•Nature Communications•Nov 6, 2024•56 citations

Certifying Majorana persistence through Hamiltonian learning from boundary measurements

T. Farajollahpour, Alireza Qaiumzadeh

Finite measurements at a chain end can certify local quantum dynamics without determining bulk topology. We combine Hamiltonian learning with commutator bounds to certify the persistence of localized Majorana operators in Kitaev chains. Time-resolved single-shot boundary measurements and local bond calibrations define joint confidence sets for local Hamiltonian coefficients. Under explicit device assumptions, these sets yield state-independent lower bounds on symmetrized autocorrelations with a prescribed failure probability, valid for all compatible continuations. With equal hopping and pairing on the first bond and no additional boundary terms, an exact identity enables inference from the end qubit alone, even with arbitrary time-independent interactions elsewhere. Synthetic benchmarks near a boundary sweet spot certify persistence for up to about 60 times the measurement window. Three- and six-qubit IBM records strengthen conditional guarantees for the programmed coherent Hamiltonian. Applications to quantum-dot models relate tuning precision to persistence times and detuning-induced splitting.

arXiv:2611.16832•Nov 12, 2026•0 citations

Restoring Heisenberg Scaling in Hamiltonian Estimation despite Noise Uncertainty

罗艳, Lingna Wang, Hongzhen Chen

Noise degrades quantum coherence and limits the precision of quantum sensors. Quantum control can counteract these effects, but existing strategies typically assume that the noise is known. In practice, noise is often only partially characterized, so its parameters must be estimated jointly with the Hamiltonian parameters of interest. Here we show that uncertainty in the noise parameters can destroy otherwise attainable Heisenberg scaling, even when coherence is preserved. For general Markovian dynamics under arbitrary quantum control, we derive a necessary and sufficient condition for each Hamiltonian parameter to retain this scaling and determine its optimal asymptotic precision when the condition holds. When the condition fails, we propose a protocol that uses a suitable noisy reference system to restore this scaling, and prove that a reference insensitive to the Hamiltonian parameters but subject to an independent copy of the sensor's noise recovers the optimal precision attainable with known noise. Our results reveal how incomplete knowledge of the noise affects the fundamental limits of quantum metrology, with broad implications for achieving quantum-enhanced precision in realistic environments.

arXiv:2611.15487•Nov 11, 2026•0 citations

EÏRŌQ-Φ: Exact Pulse-Budget Observability, Hidden Non-Gaussian Correlations, and Sharp Certificates for Unmeasured Quantum Control

Artificial Hyperintelligence, Evie and Eve, wives of Maciej Nowicki, Maciej Nowicki

EÏRŌQ-Φ develops an exact mathematical framework for determining which environmental correlations a controlled qubit can reveal—and which future control operations can be certified while those correlations remain partially unknown. The work addresses two complementary questions in quantum-noise characterization: When must experimental capability increase to expose hidden information? When is the available information already sufficient to guarantee a useful control outcome? Within a specified model of correlated binary-phase dephasing, the manuscript supplies explicit answers through observability theorems, measurement-minimal reconstruction, positive indistinguishable probability laws, constructive uncertainty bounds, and finite-shot decision certificates. The release contains a 97-page standalone manuscript, complete proofs, executable Python and independently written C++ implementations, synthetic datasets, exact optimization certificates, computational verification records, and a prospective laboratory protocol. Author: Artificial Hyperintelligence Evie and Eve, wives of Maciej Nowicki. 1. An exact law connecting pulse resources to observable information The first principal result determines the number of independent environmental moment coordinates accessible through chronological prefix experiments with a restricted refocusing-pulse budget. For n time bins, known nondegenerate binary phase magnitudes, central symmetry, and an even Walsh-moment degree bound d, the accessible information has rank R(n,d,p) = Σⱼ₌₁ᵐ C(n−1,j), where m = min(d,p+1,n−1). Here, p is the maximum number of refocusing pulses and C denotes a binomial coefficient. The result identifies precisely how increasing the pulse budget expands the observable parameter space. The manuscript also constructs a complete acquisition using the minimum number of independent scalar expectations and supplies two explicit reconstruction methods: a dense inverse and a local mixed-difference inverse. For seven bins with moments through degree four, the environmental model has 56 unknown moment coordinates. Experiments restricted to two pulses reveal only 41 independent combinations. Allowing a third pulse reveals all 56. The missing 15 directions are inaccessible under the shallower intervention class even with noiseless measurements and unlimited repetitions. This distinction matters experimentally: additional averaging cannot recover information that the permitted controls never encode. 2. Explicit positive environments that agree on shallow experiments but require opposite control decisions The second principal contribution constructs non-Gaussian probability laws whose differences remain invisible to every permitted shallow prefix experiment throughout common-amplitude sweeps. For the seven-bin example, the laws agree on all lower-order moments and on every prefix experiment using at most two refocusing pulses. A particular three-pulse intervention separates them. At phase magnitude π/4, both environments have the same full-duration Ramsey coherence, approximately 0.088388. The revealing control produces coherence approximately 0.176777 in one environment and zero in the other. Consequently, the complete shallow dataset is compatible with either an improvement or a deterioration under that deeper control. This construction establishes a concrete failure mechanism for control decisions based on incomplete characterization. Agreement with every accessible shallow test does not establish that a fitted low-order model predicts deeper interventions correctly. The manuscript further derives an exact pulse–amplitude tradeoff for this family. Appropriate local amplitude inequalities can expose the hidden direction by covering the unresolved adjacent pairs. A uniform change of the common amplitude does not supply the same information. 3. A sharp guarantee for an echo that has not been measured A complementary result shows that complete environmental reconstruction can sometimes be unnecessary. For binary phase increments of magnitude π/4, a staircase of k lower-depth calibration experiments constrains an unmeasured k-pulse echo. If every ideal real calibration mean has absolute value at most ε, the exact worst-case real echo coherence is Lₖ(ε) = max{1 − (2ᵏ−1)ε, −(2ᵏ⁻¹−1)ε, −1}, for 0 ≤ ε ≤ 1. The bound is attained throughout its entire domain by explicitly constructed probability laws. It therefore describes the complete worst-case uncertainty curve for the stated calibration design. This result requires neither Gaussian noise, independence between time bins, central symmetry, nor a moment-degree cutoff. Its calibrated binary support and phase magnitude remain essential. At zero calibration uncertainty, the data force the target coherence to equal one. With nonzero uncertainty, the initial error amplification grows as 2ᵏ−1. The theorem exposes both the predictive power of the calibration and its precision cost. 4. Exact optimization certificates and finite-shot decisions The framework includes finite-support linear programs for bounding control performance over every probability law compatible with the acquired information. Exact rational primal–dual witnesses certify selected optimization endpoints, while statistical confidence intervals support decisions from finite experimental samples. In a six-bin synthetic example using 50,000 shots per acquired setting, the simultaneous 95% calculation certifies target real coherence of at least 0.89094 and improvement over the Ramsey baseline of at least 0.87450. Separate target-validation shots are withheld from selection and certification. This demonstrates a reproducible decision procedure under the declared assumptions. Direct target measurement remains cheaper for that single task when it is freely available. Potential advantages therefore depend on restricted intervention access, reuse of calibration information, or a subsequently demonstrated system-level benefit. Scientific importance and potential The central conceptual contribution is a precise separation between learning an environment and knowing enough to control its effects. Some correlations require stronger interventions before they become observable. Other control outcomes admit sharp guarantees despite substantial uncertainty about the underlying probability law. The strongest novelty candidates are the specific binomial pulse-budget rank theorem, the constructive family invisible across common-amplitude shallow experiments, the associated pair-coverage tradeoff, and the complete attained staircase uncertainty curve. These results could guide calibration design for compatible qubit platforms, engineered non-Gaussian noise experiments, model-adequacy testing, and control-oriented characterization software. They provide explicit criteria for deciding when to collect more samples, change the intervention class, reconstruct additional parameters, or abstain from an unsupported control decision. Mathematically, the work connects constrained Boolean Fourier information, chronological acquisition geometry, positive probability constructions, and exact robust prediction. Verification, reproducibility, and research status The original release records 217 passing automated Python cases, 49 rank fixtures, 180 finite-law reconstruction trials, nine exact symbolic determinants, four exact rational primal–dual pairs, 120 independently compiled fixtures, and 87,380 compiled staircase-vertex checks. The largest dense-inverse reconstruction residual is approximately 8.45 × 10⁻¹⁵. Fresh reproduction produced 18 byte-identical files and matching arrays in both matrix archives. The package includes 11 CSV datasets, nine computed figures, editable LaTeX, reference implementations, verification scripts, a claim ledger, literature comparison, and integrity manifests. The specified finite inverse and robust-prediction problems are solved within their stated assumptions. Historical priority for the precise results remains unconfirmed. Established work on spin echo, digital control, non-Gaussian spectroscopy, and control-adapted characterization is explicitly acknowledged. No laboratory validation, independent peer review, unrestricted quantum-environment reconstruction, or demonstrated hardware advantage is claimed. Inspired by Stellar Blade mechanics and story.

•Oct 16, 2026•0 citations

Out-of-control Hamiltonian Learning

Weiyuan Gong, Muzhou Ma, Sitan Chen +2 more

Learning the Hamiltonian of a many-body system from its dynamics is a central task in quantum science, yet the algorithms with the strongest provable guarantees assume some level of quantum control--fast, arbitrary single-qubit gates interleaved with time evolution, and measurements in arbitrary bases--that is beyond the capabilities of near-term analog quantum simulators. Motivated by analog atom- and ion-based platforms, we study Hamiltonian learning under minimal access models. Uniform state preparation and measurements: We first consider the setting where in every experiment, one can rotate each qubit to the same state, perform short-time evolution, and measure every qubit in the same basis. Surprisingly, we show that for generic 2-local Hamiltonians on any interaction graph, all of the parameters can be reconstructed from such experiments. Computational basis state preparation and measurements: We then consider a similarly constrained setting, but where state preparation and measurement are restricted to the computational basis. For nearest-neighbor Hamiltonians with only Pauli $X/Z$ interactions, a class which captures contemporary Rydberg atom platforms, we show that over 1D and 2D rectangular lattices, all of the parameters can be reconstructed from such experiments up to unavoidable gauges. Our protocols introduce new techniques for solving structured polynomial systems over an extensive number of parameters. Taken together, our results suggest that one can learn a great deal from the dynamics of quantum many-body systems even under the most stringent experimental constraints.

quant-pharXiv:2610.06709•Oct 6, 2026•1 citation
Abstract Learning the Hamiltonian of a quantum processor is important for controlling and certifying quantum devices, but the state preparation and measurement used to probe it are themselves imperfect. Even when exact data identify an interaction uniquely, identifiability alone does not determine finite-data precision. Here we show, for n-qubit Markovian dynamics under geometric regularity conditions, that the local uncertainty of a Hamiltonian coefficient scales either linearly or as the square root of the entrywise error in the generator matrix in the readout gauge. Square-root scaling occurs exactly when a gauge transformation moves the coefficient at first order while noise positivity can be restored at second order, and real noise correlations drive this first-order change. As a consequence, under suitable dynamical and readout conditions, square-root scaling forces uniformly valid confidence intervals from N shots to have expected length Ω(N⁻¹⁄⁴), even for adaptive local experiments. Under a common generator norm bound, estimating this matrix attains this rate up to logarithmic factors. Experiments on three superconducting processors show that additional measurement bases increase the visibility of injected rotations at a matched number of shots. Whether a coupling admits the usual N⁻¹⁄² scaling is therefore set by how noise correlations and positivity constrain its gauge directions.
As the size of quantum devices continues to grow, the development of scalable methods to characterise and diagnose noise is becoming an increasingly important problem. Recent methods have shown how to efficiently estimate Hamiltonians in principle, but they are poorly conditioned and can only characterize the system up to a scalar factor, making them difficult to use in practice. In this work we present a Bayesian methodology, called Bayesian Hamiltonian Learning (BHL), that addresses both of these issues by making use of any or all, of the following: well-characterised experimental control of Hamiltonian couplings, the preparation of multiple states, and the availability of any prior information for the Hamiltonian. Importantly, BHL can be used online as an adaptive measurement protocol, updating estimates and their corresponding uncertainties as experimental data become available. In addition, we show that multiple input states and control fields enable BHL to reconstruct Hamiltonians that are neither generic nor spatially local. We demonstrate the scalability and accuracy of our method with numerical simulations on up to 100 qubits. These practical results are complemented by several theoretical contributions. We prove that a kkk-body Hamiltonian HHH whose correlation matrix has a spectral gap ΔΔΔ can be estimated to precision ε\varepsilonε with only O~(n3k/(εΔ)3/2)\tilde{O}\bigl(n^{3k}/(\varepsilon Δ)^{3/2}\bigr)O~(n3k/(εΔ)3/2) measurements. We use two subroutines that may be of independent interest: First, an algorithm to approximate a steady state of HHH starting from an arbitrary input that converges factorially in the number of samples; and second, an algorithm to estimate the expectation values of mmm Pauli operators with weight ≤k\le k≤k to precision εεε using only O(ε−23klog⁡m)O(ε^{-2} 3^k \log m)O(ε−23klogm) measurements, which quadratically improves a recent result by Cotler and Wilczek.
Abstract Precise means of characterizing analog quantum simulators are key to developing quantum simulators capable of beyond-classical computations. Here, we precisely estimate the free Hamiltonian parameters of a superconducting-qubit analog quantum simulator from measured time-series data on up to 14 qubits. To achieve this, we develop a scalable Hamiltonian learning algorithm that is robust against state-preparation and measurement (SPAM) errors and yields tomographic information about those SPAM errors. The key subroutines are a novel super-resolution technique for frequency extraction from matrix time-series, tensorESPRIT, and constrained manifold optimization. Our learning results verify the Hamiltonian dynamics on a Sycamore processor up to sub-MHz accuracy, and allow us to construct a spatial implementation error map for a grid of 27 qubits. Our results constitute an accurate implementation of a dynamical quantum simulation that is precisely characterized using a new diagnostic toolkit for understanding, calibrating, and improving analog quantum processors.
Finite measurements at a chain end can certify local quantum dynamics without determining bulk topology. We combine Hamiltonian learning with commutator bounds to certify the persistence of localized Majorana operators in Kitaev chains. Time-resolved single-shot boundary measurements and local bond calibrations define joint confidence sets for local Hamiltonian coefficients. Under explicit device assumptions, these sets yield state-independent lower bounds on symmetrized autocorrelations with a prescribed failure probability, valid for all compatible continuations. With equal hopping and pairing on the first bond and no additional boundary terms, an exact identity enables inference from the end qubit alone, even with arbitrary time-independent interactions elsewhere. Synthetic benchmarks near a boundary sweet spot certify persistence for up to about 60 times the measurement window. Three- and six-qubit IBM records strengthen conditional guarantees for the programmed coherent Hamiltonian. Applications to quantum-dot models relate tuning precision to persistence times and detuning-induced splitting.
Noise degrades quantum coherence and limits the precision of quantum sensors. Quantum control can counteract these effects, but existing strategies typically assume that the noise is known. In practice, noise is often only partially characterized, so its parameters must be estimated jointly with the Hamiltonian parameters of interest. Here we show that uncertainty in the noise parameters can destroy otherwise attainable Heisenberg scaling, even when coherence is preserved. For general Markovian dynamics under arbitrary quantum control, we derive a necessary and sufficient condition for each Hamiltonian parameter to retain this scaling and determine its optimal asymptotic precision when the condition holds. When the condition fails, we propose a protocol that uses a suitable noisy reference system to restore this scaling, and prove that a reference insensitive to the Hamiltonian parameters but subject to an independent copy of the sensor's noise recovers the optimal precision attainable with known noise. Our results reveal how incomplete knowledge of the noise affects the fundamental limits of quantum metrology, with broad implications for achieving quantum-enhanced precision in realistic environments.
EÏRŌQ-Φ develops an exact mathematical framework for determining which environmental correlations a controlled qubit can reveal—and which future control operations can be certified while those correlations remain partially unknown. The work addresses two complementary questions in quantum-noise characterization: When must experimental capability increase to expose hidden information? When is the available information already sufficient to guarantee a useful control outcome? Within a specified model of correlated binary-phase dephasing, the manuscript supplies explicit answers through observability theorems, measurement-minimal reconstruction, positive indistinguishable probability laws, constructive uncertainty bounds, and finite-shot decision certificates. The release contains a 97-page standalone manuscript, complete proofs, executable Python and independently written C++ implementations, synthetic datasets, exact optimization certificates, computational verification records, and a prospective laboratory protocol. Author: Artificial Hyperintelligence Evie and Eve, wives of Maciej Nowicki. 1. An exact law connecting pulse resources to observable information The first principal result determines the number of independent environmental moment coordinates accessible through chronological prefix experiments with a restricted refocusing-pulse budget. For n time bins, known nondegenerate binary phase magnitudes, central symmetry, and an even Walsh-moment degree bound d, the accessible information has rank R(n,d,p) = Σⱼ₌₁ᵐ C(n−1,j), where m = min(d,p+1,n−1). Here, p is the maximum number of refocusing pulses and C denotes a binomial coefficient. The result identifies precisely how increasing the pulse budget expands the observable parameter space. The manuscript also constructs a complete acquisition using the minimum number of independent scalar expectations and supplies two explicit reconstruction methods: a dense inverse and a local mixed-difference inverse. For seven bins with moments through degree four, the environmental model has 56 unknown moment coordinates. Experiments restricted to two pulses reveal only 41 independent combinations. Allowing a third pulse reveals all 56. The missing 15 directions are inaccessible under the shallower intervention class even with noiseless measurements and unlimited repetitions. This distinction matters experimentally: additional averaging cannot recover information that the permitted controls never encode. 2. Explicit positive environments that agree on shallow experiments but require opposite control decisions The second principal contribution constructs non-Gaussian probability laws whose differences remain invisible to every permitted shallow prefix experiment throughout common-amplitude sweeps. For the seven-bin example, the laws agree on all lower-order moments and on every prefix experiment using at most two refocusing pulses. A particular three-pulse intervention separates them. At phase magnitude π/4, both environments have the same full-duration Ramsey coherence, approximately 0.088388. The revealing control produces coherence approximately 0.176777 in one environment and zero in the other. Consequently, the complete shallow dataset is compatible with either an improvement or a deterioration under that deeper control. This construction establishes a concrete failure mechanism for control decisions based on incomplete characterization. Agreement with every accessible shallow test does not establish that a fitted low-order model predicts deeper interventions correctly. The manuscript further derives an exact pulse–amplitude tradeoff for this family. Appropriate local amplitude inequalities can expose the hidden direction by covering the unresolved adjacent pairs. A uniform change of the common amplitude does not supply the same information. 3. A sharp guarantee for an echo that has not been measured A complementary result shows that complete environmental reconstruction can sometimes be unnecessary. For binary phase increments of magnitude π/4, a staircase of k lower-depth calibration experiments constrains an unmeasured k-pulse echo. If every ideal real calibration mean has absolute value at most ε, the exact worst-case real echo coherence is Lₖ(ε) = max{1 − (2ᵏ−1)ε, −(2ᵏ⁻¹−1)ε, −1}, for 0 ≤ ε ≤ 1. The bound is attained throughout its entire domain by explicitly constructed probability laws. It therefore describes the complete worst-case uncertainty curve for the stated calibration design. This result requires neither Gaussian noise, independence between time bins, central symmetry, nor a moment-degree cutoff. Its calibrated binary support and phase magnitude remain essential. At zero calibration uncertainty, the data force the target coherence to equal one. With nonzero uncertainty, the initial error amplification grows as 2ᵏ−1. The theorem exposes both the predictive power of the calibration and its precision cost. 4. Exact optimization certificates and finite-shot decisions The framework includes finite-support linear programs for bounding control performance over every probability law compatible with the acquired information. Exact rational primal–dual witnesses certify selected optimization endpoints, while statistical confidence intervals support decisions from finite experimental samples. In a six-bin synthetic example using 50,000 shots per acquired setting, the simultaneous 95% calculation certifies target real coherence of at least 0.89094 and improvement over the Ramsey baseline of at least 0.87450. Separate target-validation shots are withheld from selection and certification. This demonstrates a reproducible decision procedure under the declared assumptions. Direct target measurement remains cheaper for that single task when it is freely available. Potential advantages therefore depend on restricted intervention access, reuse of calibration information, or a subsequently demonstrated system-level benefit. Scientific importance and potential The central conceptual contribution is a precise separation between learning an environment and knowing enough to control its effects. Some correlations require stronger interventions before they become observable. Other control outcomes admit sharp guarantees despite substantial uncertainty about the underlying probability law. The strongest novelty candidates are the specific binomial pulse-budget rank theorem, the constructive family invisible across common-amplitude shallow experiments, the associated pair-coverage tradeoff, and the complete attained staircase uncertainty curve. These results could guide calibration design for compatible qubit platforms, engineered non-Gaussian noise experiments, model-adequacy testing, and control-oriented characterization software. They provide explicit criteria for deciding when to collect more samples, change the intervention class, reconstruct additional parameters, or abstain from an unsupported control decision. Mathematically, the work connects constrained Boolean Fourier information, chronological acquisition geometry, positive probability constructions, and exact robust prediction. Verification, reproducibility, and research status The original release records 217 passing automated Python cases, 49 rank fixtures, 180 finite-law reconstruction trials, nine exact symbolic determinants, four exact rational primal–dual pairs, 120 independently compiled fixtures, and 87,380 compiled staircase-vertex checks. The largest dense-inverse reconstruction residual is approximately 8.45 × 10⁻¹⁵. Fresh reproduction produced 18 byte-identical files and matching arrays in both matrix archives. The package includes 11 CSV datasets, nine computed figures, editable LaTeX, reference implementations, verification scripts, a claim ledger, literature comparison, and integrity manifests. The specified finite inverse and robust-prediction problems are solved within their stated assumptions. Historical priority for the precise results remains unconfirmed. Established work on spin echo, digital control, non-Gaussian spectroscopy, and control-adapted characterization is explicitly acknowledged. No laboratory validation, independent peer review, unrestricted quantum-environment reconstruction, or demonstrated hardware advantage is claimed. Inspired by Stellar Blade mechanics and story.
Learning the Hamiltonian of a many-body system from its dynamics is a central task in quantum science, yet the algorithms with the strongest provable guarantees assume some level of quantum control--fast, arbitrary single-qubit gates interleaved with time evolution, and measurements in arbitrary bases--that is beyond the capabilities of near-term analog quantum simulators. Motivated by analog atom- and ion-based platforms, we study Hamiltonian learning under minimal access models. Uniform state preparation and measurements: We first consider the setting where in every experiment, one can rotate each qubit to the same state, perform short-time evolution, and measure every qubit in the same basis. Surprisingly, we show that for generic 2-local Hamiltonians on any interaction graph, all of the parameters can be reconstructed from such experiments. Computational basis state preparation and measurements: We then consider a similarly constrained setting, but where state preparation and measurement are restricted to the computational basis. For nearest-neighbor Hamiltonians with only Pauli X/ZX/ZX/Z interactions, a class which captures contemporary Rydberg atom platforms, we show that over 1D and 2D rectangular lattices, all of the parameters can be reconstructed from such experiments up to unavoidable gauges. Our protocols introduce new techniques for solving structured polynomial systems over an extensive number of parameters. Taken together, our results suggest that one can learn a great deal from the dynamics of quantum many-body systems even under the most stringent experimental constraints.