{"id":"317075fd-3b1e-4ca3-9462-fc01fdc990c2","arxiv_id":"2508.10593","paper_version":1,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":1,"one_line_summary":"A VQE ansatz built from entropy and mutual information reaches 99.99% accuracy with up to 99% fewer two-qubit gates than unitary coupled cluster for atoms.","lead":"This paper proposes a new way to design quantum circuits for atomic physics simulations, using quantum correlation measures to decide where to place entanglers. If it works, it could make quantum chemistry calculations on near-term quantum computers far more efficient.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Target-state fidelity is the linchpin: the abstract gives no evidence that the approximate multi-qubit state used for entropy/mutual-information calculations faithfully captures the true ground-state correlations.","rationale":"The reader identified the approximate target state as the weakest assumption, and I agree. The central claim—that the quantum-information-inspired ansatz achieves chemical accuracy with far fewer two-qubit gates than UCC—depends critically on the target state's ability to reveal the correct entanglement structure of the true ground state. The abstract offers no detail on how the target state is obtained or validated, and the performance could be circular if the target state is already near-exact. My concrete test would settle whether the method is robust to the target-state choice or requires a target state that already encodes the relevant correlations. Since the full manuscript is not available, I cannot assess whether the paper already contains such a test; therefore the appropriate verdict is CONDITIONAL: the claim should be accepted only if the target-state robustness is demonstrated or if the target state is shown to be a faithful proxy with independent justification. This does not reject the paper's premise but flags the one assumption that, if unwarranted, would invalidate the central numerical results.","tokens_in":731,"tokens_out":2739,"duration_ms":30830,"concrete_test":"For a small system explicitly treated in the paper (e.g., Be or LiH with 4–12 qubits), re-run the ansatz construction using two very different approximate target states: (a) a restricted Hartree–Fock determinant and (b) a configuration interaction singles and doubles (CISD) wavefunction. Compare the resulting VQE energies against the reported CASCI values. If the final energy is insensitive to the target state's quality (i.e., both give ≥99.99% accuracy), the method is robust. If the accuracy drops sharply when the Hartree–Fock target state is used, then the method relies on a target state that already contains the essential correlations, and the abstract's claim cannot stand without specifying and justifying that target state.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The construction's entire logic is: (1) start with an approximate multi-qubit target state, (2) compute von Neumann entropies and mutual information from it, (3) place two-qubit entanglers on the pairs with maximum correlations, (4) variationally optimize the resulting circuit. Step (3) is only as good as step (1). If the target state is too crude—say, a single Slater determinant or a state that misses static correlation—its mutual information will be structurally different from that of the true ground state. For example, a mean-field state has zero mutual information between orbitals with opposite spin in a restricted formalism, so the algorithm would place no entanglers where static correlation is essential. The abstract does not specify the provenance of the target state (e.g., HF, CISD, DMRG, or a small CASCI) nor provide any evidence that its correlation profile matches the exact CASCI state for the tested atoms. Without this, the reported 99.99% accuracy could be an artifact of choosing a target state that already contains the relevant correlations, making the ansatz construction circular rather than predictive. This is the weakest load-bearing assumption because if it fails, the entire method collapses to a random or misleading entangler placement, and the claimed gate reduction becomes irrelevant. A secondary concern is the fairness of the '99% fewer 2-qubit gates' comparison to UCC, but that is subordinate to the correctness of the ansatz itself.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript proposes a quantum information-inspired ansatz for VQE, in which an approximate multi-qubit target state is used to compute von Neumann entropies and quantum mutual information. These quantities guide the deterministic placement of two-qubit entanglers in a parametrized quantum circuit. The authors claim that for atomic systems with up to 12 qubits (12 spin orbitals), the ansatz yields ground-state energies with 99.99% accuracy relative to complete active space configuration interaction (CASCI), while using two blocks that contain at most 99% fewer two-qubit gates than the UCC ansatz. This assessment is based solely on the abstract; the full text was not available for review.","tokens_in":1073,"tokens_out":2881,"duration_ms":31525,"significance":"If the claims are substantiated, the approach addresses a key bottleneck of VQE—circuit depth—by using quantum information-theoretic measures to guide ansatz construction. The idea of replacing heuristic entangler placements with correlation-based metrics is promising and could be a step toward practical quantum chemistry simulations. However, the abstract alone provides no evidence for these claims, and the central assumption that an approximate target state faithfully captures the true ground-state correlations is unverified. The significance therefore remains conditional on a full technical report that does not currently exist in the provided material.","major_comments":[{"comment":"The construction's linchpin is the 'approximate multi-qubit target state' used to compute entropies and mutual information. The abstract does not specify how this state is obtained (e.g., Hartree–Fock, CISD, DMRG, or a small CASCI) or how well it approximates the CASCI ground state for the tested atoms. If the target state misses static correlation, the mutual information will misidentify the relevant qubit pairs, and the entangler placement will be suboptimal. A rigorous analysis of target-state sensitivity is necessary to establish the method's validity.","section":"Abstract"},{"comment":"The claim of '99.99% accuracy relative to the complete active space configuration interaction values' is not supported by any numerical data. The abstract reports no energies, error bars, or system specifications; the metric 'accuracy' is undefined (percent error in energy? state fidelity?). The full text must include tables or plots of computed energies for all tested atomic systems, along with convergence details, to make this claim falsifiable.","section":"Abstract"},{"comment":"The gate-reduction comparison with UCC is ambiguous. The abstract states 'at most 99% fewer 2-qubit gates', which could mean the reduction factor varies and can be as low as 0% (i.e., 'at most' 99%), undercutting the claim. More importantly, the number of two-qubit gates depends on the choice of UCC variant (e.g., UCCSD, UpCCGSD) and on the specifics of the two blocks. Without a precise definition of the ansatz and the baseline, the reduction claim cannot be evaluated. The full text must define the circuits and count gates explicitly.","section":"Abstract"},{"comment":"Scalability beyond 12 spin orbitals is not addressed. The algorithm requires computing all pairwise mutual informations from the target state, and the classical pre-processing cost as well as the required fidelity of the target state for larger systems are unknown. The abstract's claims are limited to a small qubit count; the full text should include a complexity analysis or at least a discussion of scalability.","section":"Abstract"}],"minor_comments":[{"comment":"The abstract does not list the specific atomic systems studied, so the reader cannot judge the diversity of correlation regimes (e.g., weakly vs. strongly correlated systems) or the generality of the conclusions.","section":"Abstract"},{"comment":"The phrase 'at most 99% fewer' is likely a typo for 'at least 99% fewer' if the intended meaning is a large reduction. The wording currently suggests an upper bound on the reduction.","section":"Abstract"},{"comment":"The term 'unique blocks' is not defined in the abstract. It would be helpful to state whether the two blocks are fixed or optimized during the VQE, and how the entangler placement within each block is parameterized.","section":"Abstract"}],"recommendation":"uncertain","confidential_remarks":"This review is based only on the abstract. The paper's central claims cannot be verified without the full text, which was not provided. The editor may wish to request the complete manuscript or clarify the review policy for abstract-only submissions. If the full text contains the missing derivations, numerical data, and an explicit analysis of the target-state assumption, the paper could merit further consideration; as it stands, I cannot assess its soundness."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You asked for my take on this one. I'm working from the abstract only, so treat everything as provisional.\n\nWhat's actually new: using von Neumann entropy and quantum mutual information computed from an approximate multi-qubit target state to place two-qubit entanglers deterministically is a reasonable idea, and distinct from the usual ADAPT-VQE style of picking operators by energy gradients. The claimed gate reduction—up to 99% fewer two-qubit gates than UCC while keeping energy accuracy at 99.99% relative to CASCI for atoms up to 12 qubits—is impressive if it holds. The authors also deserve credit for testing on real atomic systems rather than just toy molecules.\n\nThe soft spot is exactly the one the stress-test note flags: we don't know where the target state comes from. If it's a single Slater determinant, the mutual information will miss static correlation entirely, and the selected entanglers will be misdirected. If it's a small CASCI or CISD, then the method is partly circular—you're using a state that already contains the correlations you're trying to discover. The abstract says 'approximate multi-qubit target state' but doesn't specify its provenance. That's not a fatal flaw by itself; the variational optimization might correct some misplacements, and a cheap target state like CISD might be good enough. But the authors need to show the sensitivity of the final energy to the choice of target state, or at least report what they used.\n\nA secondary concern is the gate-count comparison. 'At most 99% fewer' is a strange way to phrase it—it sounds like they're reporting the best case, not the typical case. And the comparison to UCC should include the number of parameters and the optimization cost, not just two-qubit gate count.\n\nThe paper deserves a serious referee. The idea is testable and the results, if verified, would be practically useful for NISQ chemistry. But I'd want the referee to push hard on the target-state question and on the fairness of the comparison to other adaptive ansätze. I'd bring it to a reading group once the full text is available, but I wouldn't cite it yet.","headline":"Plausible information-guided ansatz, but the abstract's key premise—the source and fidelity of the approximate target state—is unstated, so the circularity question remains open until we see the full method.","tokens_in":699,"tokens_out":706,"would_cite":false,"duration_ms":19448,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.Ac","31.15.-p"],"model":"deepseek-v4-flash","headline":"A quantum-information-inspired ansatz yields 99.99% accurate atomic energies with up to 99% fewer two-qubit gates than UCC.","keywords":["variational quantum eigensolver","quantum mutual information","von Neumann entropy","ansatz construction","atomic ground-state energy","circuit depth reduction","unitary coupled-cluster","two-qubit entanglers"],"falsifier":"Run the method on a strongly correlated system where a low-cost approximate state is known to miss static correlation (for example, a bond stretched well past equilibrium). If the two-block ansatz built from that proxy fails to reach close-to-exact ground-state energy while the full UCC ansatz succeeds, the information-guidance premise is refuted. Alternatively, compare the mutual-information-ranked qubit pairs against the exact reduced two-qubit density matrices: any significant mismatch between the ranking and the true largest correlations would break the construction.","tokens_in":671,"feed_emoji":"⚛️","tokens_out":4813,"duration_ms":50908,"temperature":0.7,"pith_summary":"This paper tries to show that a variational quantum eigensolver ansatz can be designed from the correlations of an approximate target state rather than by heuristic guesswork. The authors propose a concrete construction: compute von Neumann entropies and quantum mutual information from a cheap approximate multi-qubit state, rank qubit pairs by shared correlation, and place two-qubit entanglers accordingly in a shallow circuit. Applied to atomic systems with up to 12 qubits in noiseless simulation, the resulting two-block ansatz reproduces complete-active-space configuration-interaction energies to 99.99% accuracy while using at most 99% fewer two-qubit gates than the unitary coupled-cluster ansatz. If it holds up, the method gives a deterministic, resource-efficient route to VQE circuits, which matters because circuit depth is the main bottleneck on near-term quantum hardware.","feed_headline":"VQE ansatz cuts qubit gates 99% at near-exact energies","feed_subtitle":"Correlation measures, not heuristics, decide where entanglers go, shrinking circuits for atomic ground-state calculations.","key_machinery":"The mechanism is the pair of quantum-information measures computed from the approximate target state. von Neumann entropy quantifies the entanglement of each qubit with the rest of the system, and quantum mutual information between pairs quantifies how much correlation two qubits share beyond what each has with the environment. Ranking qubit pairs by mutual information tells the circuit designer which two-qubit entanglers to place, and the entropy distribution tells how to group them into blocks. This turns ansatz construction from a heuristic search into a deterministic, information-guided procedure.","core_discovery":"The central claim is that the information content of an approximate multi-qubit target state—quantified by von Neumann entropy and quantum mutual information—is enough to deterministically construct a VQE ansatz that is both accurate and shallow. Previous heuristics guess circuit structure; here the target state's correlation structure identifies which qubit pairs should be connected by two-qubit entanglers and groups them into blocks. For the atomic test cases, at most 12 qubits, two such blocks suffice to reach 99.99% agreement with complete-active-space configuration-interaction ground-state energies, while the two-qubit-gate count is at least 99% smaller than in the UCC ansatz. The paper","pith_inferences":["Editorial inference: the accuracy ceiling is set by the approximate target state, so the method should be tested with progressively cheaper proxies (e.g., mean-field vs. low-depth classical correlated states) to map when it starts to fail.","Editorial inference: the same mutual-information criterion could be applied to select entanglers adaptively during the VQE optimization, re-computing the target state from measured reduced density matrices; this would make the method self-correcting for strongly correlated cases.","Editorial inference: the information-guided placement may also transfer to other variational algorithms that use parameterized circuits, such as ground-state preparation for lattice models, as long as a decent approximate state is available.","Editorial inference: a direct head-to-head against established adaptive ansatz builders on strongly correlated molecules would clarify whether the information-theoretic ranking adds value beyond simply adding entanglers in order of decreasing correlation."],"forward_implications":["If the approach works beyond these test cases, VQE circuits for quantum chemistry can be generated from a classical approximate state instead of by trial and error, removing a major design bottleneck.","The reported 99% reduction in two-qubit gates directly lowers the circuit depth that must survive on noisy hardware, making near-term atomic and molecular simulations more feasible.","The method's accuracy is tied to the quality of the approximate target state; upgrading that initial proxy should systematically improve the chosen ansatz.","Because entangler placement is deterministic, the ansatz is reproducible and could be ported to different qubit topologies without re-optimizing the whole circuit structure.","For larger active spaces beyond 12 qubits, the same information-guided procedure may keep the circuit depth growing more slowly than UCC, provided the target state remains a good proxy."],"supporting_citations":[],"fun_headline_variants":["Info-based ansatz: 99% fewer VQE gates","Quantum correlations design shallow VQE circuits","Entropy-guided VQE ansatz: 99% gate reduction","VQE ansatz from quantum info: accurate, shallow","Entropy and mutual info slash VQE circuit depth"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The approximate multi-qubit target state must already carry the essential correlations of the true ground state; if it does not, the mutual-information ranking will place entanglers on the wrong qubit pairs and the claimed accuracy will not hold.","fun_headline_variants_meta":{"raw":{"variants":["Info-based ansatz: 99% fewer VQE gates","Quantum correlations design shallow VQE circuits","Entropy-guided VQE ansatz: 99% gate reduction","VQE ansatz from quantum info: accurate, shallow","Entropy and mutual info slash VQE circuit depth"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001489,"raw_usage":{"total_tokens":5832,"prompt_tokens":775,"completion_tokens":5057,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":519,"completion_tokens_details":{"reasoning_tokens":4978}},"tokens_in":519,"tokens_out":5057,"duration_ms":36260,"temperature":1.0,"reasoning_tokens":4978,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T20:19:19.343933+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the method on a strongly correlated system where a low-cost approximate state is known to miss static correlation (for example, a bond stretched well past equilibrium). If the two-block ansatz built from that proxy fails to reach close-to-exact ground-state energy while the full UCC ansatz succeeds, the information-guidance premise is refuted. Alternatively, compare the mutual-information-ranked qubit pairs against the exact reduced two-qubit density matrices: any significant mismatch between the ranking and the true largest correlations would break the construction.","supporting_citations":[],"review_version":1}