REVIEW 3 major objections 2 minor 1 cited by
Guided sampling ans\"atzes for variational quantum computing
T0 review · 3 major / 2 minor · reviewed 2026-08-05 · deepseek-v4-flash
Pith's one-line read The paper claims that a guided sampling ansatz—a trial state shaped by measured samples, system interactions, and a parameter space—can compute total energies for H3O+ with errors below 1.59×10^-3 Ha using only 200 circuit executions, excee
desk verdict A potentially useful VQE trick with a genuine hardware demo, but the abstract leaves an unresolved question about whether the same measured samples are used to shape the ansatz and to compute the energy. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The central object is the guided sampling ansatz (GSA): a parametrized trial state whose defining characteristic is that its functional form is generated from (i) the interaction terms of the simulated system; (ii) samples of measured quantum states obtained during the computation; and (iii) an associated parameter space. The measured samples serve a dual role: they inform the structure of the ansatz, restricting the variational search to a subspace that the quantum device has already shown to be occupied, while the parameter space gives the optimizer room to adjust amplitudes. This mechanism is what allows the circuit-execution count to remain low—200 per structure—because the ansatz itself
What would settle it
A single numerical experiment would settle the central claim: run the GSA optimization on H3O+ with the same 200-circuit budget but evaluate the final total energy using a fresh set of measurement samples that were never used to guide the ansatz. If the error rises above 1.59e-3 Ha (i.e., chemical accuracy is lost), the reported sub-chemical-accuracy result is a fitting artifact of sample reuse; if the error stays below the threshold, the claim of measurement-guided accuracy is confirmed.
Extended reading notes
Core claim
The central claim is that a guided sampling ansatz—a parametrized trial state whose functional form depends on the interaction terms of the target Hamiltonian, the measured state samples collected during the run, and a modest parameter space—can deliver sub-chemical-accuracy total energies with an extremely small number of circuit executions. The demonstration is carried out on the hydronium cation H3O+ using a minimal GSA on a trapped-ion quantum computer (IonQ Aria). With only 200 circuit executions per structure, the computed total energies around the relaxed structure have errors below 1.59×10^-3 Ha, which the authors equate with exceeding chemical accuracy. The discovery, if correct, is
Load-bearing premise
The load-bearing premise is that the measured state samples used to build the guided sampling ansatz do not also bias the energy estimate; if the same samples that pick the ansatz form are reused to compute the total energy, the reported sub-chemical-accuracy errors could be a fitting artifact rather than a predictive result.
Editorial extensions
If this is right
- If GSAs behave as demonstrated, variational quantum eigensolvers can reach chemically accurate total energies with a few hundred circuit executions per structure, removing a major practical bottleneck of current hardware.
- The ansatz is selected partly by measured samples, so the method is not tied to a specific molecule; the same guiding principle should transfer to other systems whose interactions are known.
- A minimal GSA already suffices for a charged molecular ion like H3O+, suggesting that small ansätze can be effective when they are data-guided rather than pre-fixed.
- The explicit dependence on system interactions means the ansatz can, in principle, be generated automatically from the Hamiltonian, which would eliminate much of the manual ansatz design in quantum chemistry.
- The modest parameter space combined with 200 circuit executions hints that the method could be tested systematically on larger molecules, though the abstract does not report scaling data.
Reading between the lines
- The most direct test the paper leaves implicit is sample separation: because measured samples shape the ansatz and measurements also estimate the energy, a skeptic would want to see a validation set of fresh samples. Our inference is that the 1.59e-3 Ha error may partially reflect reuse of the guiding samples in the energy estimate; a clean test is to evaluate with independent samples.
- The mechanism suggests a natural connection to adaptive ansatz techniques such as ADAPT-VQE, where operators are added based on gradients; GSA instead uses measurement populations, which could be more directly hardware-friendly, but the paper does not make that comparison.
- A further extension, drawn from the structure of GSA, would be to use the measured samples to construct a basis of 'visited' configurations and then perform a classical diagonalization in that subspace (a hybrid quantum-classical approach); this is not proposed in the abstract, but it follows from the idea that the samples define a relevant subspace.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The manuscript introduces a class of guided sampling ansätze (GSAs) for variational quantum computing, where the ansatz depends on the system interactions, measured state samples, and a parameter space. As a demonstration, the authors apply a minimal GSA to the hydronium cation H3O+ and report total energies around the relaxed structure with errors below 1.59×10^-3 Ha using only 200 circuit executions per structure on the IonQ Aria quantum computer, thus exceeding chemical accuracy. The central claim is that data-driven ansatz selection can make variational quantum chemistry both accurate and hardware-efficient.
Significance. If the reported result holds, the paper would be a notable contribution: it suggests that measurement-guided ansatz design can drastically reduce the required number of circuit executions while maintaining chemical accuracy, which is a key bottleneck for near-term quantum chemistry. The idea of making the ansatz depend on measured samples is nontrivial and could open a new direction in variational quantum eigensolvers. The empirical claim is falsifiable and, if properly validated, would be of high practical importance. However, the abstract-level evidence is insufficient to verify the central claim, and the methodological risk of sample reuse (training/evaluation circularity) must be explicitly addressed before the result can be accepted.
major comments (3)
- [Abstract] The GSA is said to 'depend on ... measured state samples as well as a parameter space,' and the reported error is obtained with the same 200 circuit executions per structure. If samples used to construct or select the ansatz are also used to estimate the total energy, the reported error is optimistically biased and is not a predictive quantity. The authors must state whether the 200 executions are partitioned into independent training and evaluation sets, or whether fresh measurements are used for the final energy estimate. Without such a separation, the sub-chemical-accuracy claim cannot be assessed.
- [Abstract] The abstract reports 'errors well below 1.59×10^-3 Ha' without any statistical uncertainty. With only 200 circuit executions per structure, shot noise is expected to contribute a non-negligible standard error to the measured energy. The authors should report error bars or a statistical analysis (e.g., standard deviation over repetitions, bootstrap estimates) to substantiate the claim of exceeding chemical accuracy. Otherwise the reported error may not be a reliable estimate.
- [Abstract] The abstract does not specify the reference method used to define 'error' in the total energy. For H3O+, a high-level classical reference (e.g., CCSD(T) in a given basis set) is typically used; the choice of reference determines the meaning of chemical accuracy. The authors should identify the reference level of theory and basis set so the claimed error is reproducible and verifiable.
minor comments (2)
- [Abstract] The phrase 'polynomial subset of the exponentially many possible solutions' is informal; the term 'subset' might be better replaced by 'parameterized subspace' or 'trial state family' to avoid confusion.
- [Abstract] The abstract does not mention the number of qubits or basis functions used for H3O+, which would help contextualize the 200-execution resource count.
Circularity Check
No circularity established from the abstract; potential sample-reuse risk is not evidenced.
full rationale
The analysis is limited to the abstract, which contains no equations, no detailed methods, and no explicit reuse of measured samples for both ansatz construction and energy evaluation. The claim that GSAs 'depend on the system interactions and measured state samples as well as a parameter space' is a description of the ansatz class, not a circular reduction. The abstract does not state whether the 200 circuit executions used to produce the energies are the same executions that guide the ansatz, nor does it exhibit any fitted parameter being renamed as a prediction. Without a specific quotation or equation showing that the energy estimate is a function of the same samples that selected the ansatz, the circularity cannot be exhibited. Per the hard rules, potential double-dipping is a risk but not a demonstrated circularity. Therefore, no significant circularity is found in the available text.
Assumptions & free parameters
free parameters (1)
- Ansatz variational parameters =
not reported in abstract
assumptions (3)
- domain assumption The variational principle: the expectation value of the Hamiltonian in any trial state is an upper bound on the ground-state energy.
- domain assumption Measurement samples drawn from the quantum device are representative of the prepared state, aside from shot noise and hardware errors.
- domain assumption The performance on H3O+ is representative of the method's general behavior.
Cite this review
Pith. "Pith review of Guided sampling ans\"atzes for variational quantum computing." pith.science (2026). https://pith.science/paper/4RK3OZH5
@misc{pith2026250813926,
author = {Pith},
title = {Pith review of: Guided sampling ans\"atzes for variational quantum computing},
year = {2026},
howpublished = {\url{https://pith.science/paper/4RK3OZH5}},
note = {Machine review of arXiv:2508.13926}
}
abstract
Quantum computing is a promising technology because of the ability of quantum computers to process vector spaces with dimensions that increase exponentially with the simulated system size. Extracting the solution, however, is challenging as the number of quantum gate operations and quantum circuit executions must still scale at most polynomially. Consequently, choosing a good ansatz--a polynomial subset of the exponentially many possible solutions--will be critical to maintain accuracy for larger systems. To address this challenge, we introduce a class of guided sampling ans\"atzes (GSAs) that depend on the system interactions and measured state samples as well as a parameter space. We demonstrate a minimal ansatz for the hydronium cation H$_3$O$^+$ and found that with only 200 circuit executions per structure on the IonQ Aria quantum computer, our calculations produced total energies around the relaxed structure with errors well below $1.59\times10^{-3}$ Ha, thus exceeding chemical accuracy.
Forward citations
Cited by 1 Pith paper
-
Hybrid VQE-CVQE algorithm using diabatic state preparation
A hybrid VQE-CVQE scheme using a few-step 'diabatic' evolution to build a guiding state, followed by classical diagonalization in the sampled subspace, yields chemically accurate ground-state energies in toy-model and...
Reviewed August 5, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.