{"id":"2c8c0b7d-6c34-431c-b354-e9dc668195c4","arxiv_id":"2607.16869","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"low","formal_verification":"none","parameter_count":3,"one_line_summary":"A Bell/Heisenberg observable needs at least three orbital-rotation contexts but one Clifford context, with an exact Eckart–Young approximation trade-off, plus a tight 2(N−1) X-rank ceiling for spin-conserving Jordan–Wigner Hamiltonians.","lead":"This paper proves exact limits on how many orbital-rotation measurement settings are needed for certain correlated observables, and shows Clifford circuits can measure some of them in a single setting. It also reports certified 31–70% shot-count savings on heavy-element molecular Hamiltonians when fully commuting Clifford settings are admitted.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No significant objection identified: the fixed-sector rank-one context lemma is internally consistent; the two-sector/number-conserving scope is explicit, and the Bell/Clifford witness checks out.","rationale":"The reader's verdict is CONDITIONAL, with the weakest assumption identified as the rank-one context model in Lemma 2. My stress-test examined exactly this premise and found it valid within the stated two-sector, number-conserving scope. The algebra of Lemma 2 follows directly from the restriction of physical orbital rotations to u_α⊗u_β on the code, and the Eckart–Young application is valid because the achievable per-context T matrices are all real rank-one 3×3 matrices. The physical Clifford witness was verified by explicit conjugation of P_X, P_Z, P_Y under Uphys. I therefore find no load-bearing mathematical flaw in the central claim. The reader's weakest-assumption wording matches the point I considered, but I do not think it lands as an objection: the number-nonconserving extension is an acknowledged open problem, and the theorem is explicitly scoped. My agreement is partial because the reader's identified assumption is indeed the load-bearing one, but I assess it as secure within scope. The remaining concerns (missing code, companion paper, dictionary caps, deferred provenance) are real but do not affect the central theorem's correctness; they support a CONDITIONAL rather than ACCEPT verdict. Hence my recommendation is UNCHANGED.","tokens_in":17194,"tokens_out":30346,"duration_ms":318221,"concrete_test":"Implement the N=2 fixed-sector model in Jordan–Wigner: for random two-sector orbital rotations (u_α,u_β) and random diagonal fragments D, compute T[(u_α⊗u_β)†D(u_α⊗u_β)] on the code subspace and verify rank≤1. Then for O_Bell=(0.7,−0.4,1.1), construct the three truncated-SVD contexts with d_i=σ_i, a_i=u_i, b_i=v_i and verify that the sum of the three fragments reproduces O_Bell on the 4D sector (Frobenius residual <1e−12), while K=1 and K=2 residuals match sqrt(σ2²+σ3²) and σ3. If any single context yields T-rank>1, or the SVD construction fails to reproduce O_Bell, Lemma 2 or Theorem 5 is wrong.","verdict_should_be":"UNCHANGED","load_bearing_attack":"I scrutinized the load-bearing premise identified by the reader: Lemma 2, which states that one particle-number-preserving orbital-rotation context contributes a rank-one two-body correlation block d_αβ ab^T. Within the paper's declared scope (two spatial orbitals per spin, independent α/β rotations, number-conserving fragments), this is sound. On the (1,1) sector the physical rotation restricts to u_α⊗u_β on the two logical qubits, and any occupation-diagonal fragment restricts to d0 I⊗I + dα Z⊗I + dβ I⊗Z + dαβ Z⊗Z. Only the Z⊗Z term contributes to T, giving exactly d_αβ ab^T. Local terms contribute to one-body Bloch vectors, not to T. The Eckart–Young step is also justified: since a and b range over all unit vectors in R^3 and d_αβ over all reals, the set of per-context T matrices is exactly all real rank-one 3×3 matrices, so K contexts cover all matrices of rank ≤K. The physical Clifford witness also checks: Uphys in Eq. (27) maps P_X→Z1, P_Z→Z3, P_Y→−Z1Z3, so one context suffices. The only potential vulnerability is the exclusion of number-nonconserving or spin-mixing Gaussian unitaries, which could in principle produce richer correlation blocks; the authors explicitly flag this as open in Remark 1 and scope the theorem to the two-sector number-conserving class. That is a stated limitation, not an internal inconsistency. No load-bearing mathematical concern identified; the remaining weaknesses are reproducibility artifacts (no shipped code/data, deferred companion framework) already noted by the reader.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper asks whether particle-number-preserving orbital rotations (fermionic Gaussian unitaries) exhaust the efficiently measurable single-context observables in quantum chemistry, compared with Clifford-accessible commuting Pauli settings. In the fixed (1,1)-particle sector of two spatial orbitals per spin, Lemma 2 shows that a single number-conserving orbital-rotation context has a two-body correlation block T = d_αβ a b^T of rank at most one. Theorem 4 uses rank subadditivity to lower-bound the required number of orbital-rotation contexts by rank T(O), and Theorem 5 identifies the best K-context approximation of T with the Eckart–Young singular-value tail. A Bell/Heisenberg witness with T = diag(λx, λy, λz) is shown to require at least three orbital-rotation contexts while one physical Clifford circuit (Eqs. 24–30) measures it. A separate result (Theorem 2) proves the X-rank ceiling r_X ≤ 2(N−1) for spin-conserving Jordan–Wigner Hamiltonians, reported tight for CH4 and NdO. Discrete RANGE search locates high-X-rank commuting families, and a companion certificate framework is used to report certified QWC versus QWC+FC shot savings of 31–70% on f-element Hamiltonians.","tokens_in":17558,"tokens_out":14356,"duration_ms":137429,"significance":"The central theoretical contribution is a clean, exact, parameter-free separation: in the fixed two-orbital sector, an orbital-rotation context contributes a rank-one correlation block, so correlation rank is a rigorous context-count lower bound, and the Eckart–Young approximation curve is exactly attainable because every unit Bloch vector and real coefficient is physically realizable. The explicit four-qubit Clifford witness is a useful operational discharge of the logical-to-physical step. The proof is self-contained and checkable. The result is deliberately scoped to the number-conserving two-orbital sector, and the authors flag generalizations as open. The numerical sections are clearly labeled as best-found or certified, and the strict separation does not depend on the search results. The X-rank ceiling is a useful diagnostic, though not by itself a Gaussian-exclusion criterion.","major_comments":[],"minor_comments":[{"comment":"The certified cost comparisons rely entirely on the companion certificate framework [1]; the certificate definition, declared caps, state model, and the claimed machine-checkable dual witnesses are not given in this manuscript. I could not verify the 31–70% numbers from the submitted material alone. Please include the certificate formalism, the dual-witness data, or a permanent artifact with the solver details. This does not affect the structural theorem, but it is needed for the advertised cost claim.","section":"Sec. VII, Table III"},{"comment":"The statement that each creation/annihilation operator contributes exactly one X/Y at its qubit is not literally correct when the same spin-orbital appears twice in a term, since the two X factors cancel in GF(2). The even-X-weight conclusion remains correct if phrased via the parity of odd-occurrence orbitals; please adjust the wording to avoid a perceptive reader finding an apparent counterexample.","section":"Sec. IV, Theorem 2 proof, Step 1"},{"comment":"The RANGE discrete mode, search settings, and witness families are promised but not included, and the continuous RANGE corroboration numbers in Sec. VB are given in text without data. For reproducibility, provide the witness families, residual traces, and scripts as supplementary material or a permanent artifact, even if in a companion release.","section":"Sec. IX / Data Availability"},{"comment":"The notation XLXL, YLYL, ZLZL is ambiguous; it should be written with spin indices, e.g. X_L^α X_L^β, and the definition of K_orbital in Eq. (19) should explicitly state that it is relative to the fixed-sector model of Sec. VA.","section":"Sec. V, Eq. (21)"},{"comment":"The caveat that term counts include the identity is important but easy to miss. Consider making the table consistent with Table I's term-count convention or adding a footnote to avoid confusion when comparing the two tables.","section":"Sec. VII A, Table II"}],"recommendation":"minor_revision","confidential_remarks":"The core algebraic result is sound and well within the standard tools of the field; I would not block acceptance on the theoretical contribution. The main risk is the unverifiable cost table from the companion manuscript, and the absence of any released code/data. I recommend requiring the companion certificate details or artifact before publication, but this is a local issue rather than a flaw in the central separation theorem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a good, honest paper. The genuinely new result is the fixed-sector correlation-rank obstruction: in the (1,1)-particle sector of two spatial orbitals per spin, one orbital-rotation context contributes a rank-one two-body correlation block T, so any observable needs at least rank T contexts, and the best K-context approximation is exactly the Eckart–Young tail. That is a clean, parameter-free statement, and the proof is sound. The Bell/Heisenberg witness with correlation rank three is a real separation: one explicit physical Clifford circuit measures the family, while three orbital-rotation contexts are necessary. The X-rank parity ceiling r_X ≤ 2(N−1) for spin-conserving Jordan-Wigner Hamiltonians is also correct as stated, and the authors are careful to call it a routing diagnostic, not a Gaussian-exclusion criterion.\n\nThe paper also does a lot of things right rhetorically. It states that Theorem 1 (Clifford simultaneous diagonalization) is standard. It flags in Remark 1 that number-nonconserving Gaussians are open. It confines the cost claims to declared capped dictionaries and describes the QWC→QWC+FC result as exactly that, not a Gaussian-versus-Clifford pricing. The companion certificate framework is referenced, but the numbers are not yet checkable.\n\nSoft spots, in proportion. The main one is reproducibility: no code or data ships with the preprint. The numerical tables and the 31–70% certified savings depend on the companion manuscript and on unstated dictionary caps, and the production f-element Hamiltonians come from an in-preparation paper. That is not a mathematical flaw, but it means a referee cannot verify the empirical claims from this document alone. The RANGE corroboration at K=3 leaves residual 1.8e-2 and does not reach zero, which is fine as a search-convergence statement, but it is weaker than the exact constructive SVD result. Minor: the Bell witness coefficients (0.7, −0.4, 1.1) are arbitrary but harmless; the rank-3 argument does not depend on them. The sector-decomposition law (Observation 1) is empirical over ten instances, and the paper says so.\n\nThe central math holds up. The right reader is someone working on measurement grouping or fermionic Gaussian measurement schemes; they should engage with the correlation-rank obstruction. It deserves a serious referee. My recommendation: send it to peer review. It needs an artifact release and closer integration of the companion framework before acceptance, but the core is worth referee time.","headline":"Solid, carefully bounded paper: the fixed-sector rank-one correlation-block lemma and Eckart–Young trade-off are genuinely new and check out; the practical savings numbers are certified but dictionary-relative and not yet independently reproducible.","tokens_in":18111,"tokens_out":2147,"would_cite":true,"duration_ms":21173,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Orbital-rotation measurement provably misses some one-Clifford observables","keywords":["orbital-rotation measurement","Clifford circuits","correlation rank","Pauli grouping","Eckart-Young approximation","X-rank","Jordan-Wigner","quantum measurement cost"],"falsifier":"Construct a single particle-number-preserving orbital-rotation context in the (1,1)-sector that reports a measured two-body correlation matrix of rank two or more; this would violate Lemma 2 and break Theorems 4-5. Short of that, measure the Heisenberg witness on the four-qubit code with two orbital-rotation contexts and check whether the residual drops below the predicted Eckart-Young tail.","tokens_in":17008,"feed_emoji":"⚛️","tokens_out":5427,"duration_ms":50262,"temperature":0.7,"pith_summary":"The paper asks whether the chemistry-standard orbital-rotation (particle-number-preserving Gaussian) approach to quantum measurement can compete with Clifford circuits for reading molecular observables. In a fixed (1,1)-particle sector of two orbitals per spin, it proves that each orbital-rotation context contributes a rank-one two-body correlation block, so any observable needs at least as many contexts as the rank of its correlation matrix, and the best K-context approximation is exactly the singular-value tail of that matrix. A Bell-diagonal Heisenberg-type observable therefore needs at least three orbital-rotation contexts, yet one explicit physical Clifford circuit measures it. This makes the two single-context measurement classes incomparable and gives a concrete structural reason to mix Gaussian and Clifford measurement settings. The paper also derives a tight parity ceiling on X-support for Jordan-Wigner Hamiltonians and shows that adding fully commuting Clifford settings cuts certified shot cost by 31-70% on four f-element systems.","feed_headline":"Bell witness needs 3 orbital rotations, 1 Clifford circuit","feed_subtitle":"A correlation-rank theorem shows the two measurement classes are provably incomparable on a Bell-type observable.","key_machinery":"The central object is the two-body correlation matrix T(O)_{ij} = 1/4 Tr[O(σ_i⊗σ_j)] of a two-qubit observable in the fixed sector. Lemma 2 shows each orbital-rotation context reports a rank-one contribution d_{αβ} a b^T, the outer product of two real Bloch vectors, so the correlation block of any K-context sum has rank at most K. The proof of the exact trade-off (Theorem 5) identifies the achievable K-context blocks with the set of real rank-K matrices and invokes the Eckart-Young theorem. The physical Bell witness uses the commuting representatives X1X2X3X4, Y1X2Y3X4, Z1Z3 and the Clifford circuit U_phys = H1 CNOT1→4 CNOT1→2 CNOT1→3, which maps them to Z strings.","core_discovery":"Within the fixed (1,1)-particle sector of two spatial orbitals per spin, a single particle-number-preserving orbital-rotation context conjugated onto an occupation-diagonal fragment produces a two-body correlation matrix of the form d a b^T, a rank-one outer product. It follows that an observable with correlation matrix T requires at least rank(T) such contexts, and the closest approximation with K contexts has residual equal to the Frobenius norm of the singular-value tail (Eckart-Young). The Bell/Heisenberg family with coefficients (λx, λy, λz) has correlation rank three; the authors exhibit four-qubit physical Pauli representatives X1X2X3X4, Y1X2Y3X4, Z1Z3 and a Clifford circuit H1 CNOT1→","pith_inferences":["If the rank-one correlation-block model extends to larger particle-number sectors, it gives a general counting obstruction: orbital-rotation dictionaries cannot serve as a drop-in replacement for Clifford measurement on correlated observables, so hybrid dictionaries are a structural necessity rather than an option.","The X-rank parity ceiling can be used as a cheap screening diagnostic: a commuting family whose X-rank approaches 2(N-1) is a strong candidate for Clifford treatment, even though rank alone does not prove Gaussian inaccessibility.","A direct hardware test would be to implement the Heisenberg witness with K=1 and K=2 orbital-rotation contexts and compare measured correlation-block residuals to the predicted singular-value tail (0.806 and 0.400 for the example coefficients); an observed residual below the target would indicate the model misses physical degrees of freedom.","The empirical sector-decomposition law (every best-found family saturates both sector projections and the deficit equals cross-sector dependencies) invites a systematic study of which molecular electronic structures produce small cross-sector dependencies, effectively classifying Hamiltonians by this measurement-relevant invariant."],"forward_implications":["Any observable in the fixed sector needs at least rank T(O) orbital-rotation contexts, and K contexts leave exactly the singular-value tail as residual; at K = rank T(O) the representation is exact.","The Bell/Heisenberg witness has correlation rank three, so the Gaussian and Clifford single-context classes are incomparable: some Clifford-accessible observables escape one orbital rotation, and some orbital-rotation contexts escape one Clifford context.","For spin-conserving Jordan-Wigner Hamiltonians, X-rank of any Pauli subset is bounded by 2(N-1), and CH4/STO-3G attains the bound with a commuting family; NdO does the same at production scale.","Controlled-Pauli insertions in Hadamard tests are Clifford, so transition-element measurement circuits carry zero T gates per execution, while sampling and state-preparation costs remain unchanged.","Declared dictionaries that enlarge product (QWC) settings with fully commuting Clifford settings certify a 31-70% shot-cost saving on four 29-35-qubit f-element Hamiltonians."],"fun_headline_variants":["Bell witness: 3 orbital rotations, 1 Clifford circuit","Rank-3 Bell witness: 3 orbital contexts, 1 Clifford","Clifford needs 1 circuit for Bell; orbital needs 3","For Bell observer: 3 orbital ops, 1 Clifford does it"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The proof treats every orbital-rotation context as a local unitary (u_α⊗u_β) acting on a diagonal occupation fragment, so the two-body block of one context is exactly a rank-one outer product d a b^T; if physical orbital rotations can report multi-fragment or number-nonconserving correlation blocks outside this form, the lower bound and approximation curve change.","fun_headline_variants_meta":{"raw":{"variants":["Bell witness: 3 orbital rotations, 1 Clifford circuit","Rank-3 Bell witness: 3 orbital contexts, 1 Clifford","Clifford needs 1 circuit for Bell; orbital needs 3","For Bell observer: 3 orbital ops, 1 Clifford does it"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00074,"raw_usage":{"total_tokens":3218,"prompt_tokens":901,"completion_tokens":2317,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":645,"completion_tokens_details":{"reasoning_tokens":2240}},"tokens_in":645,"tokens_out":2317,"duration_ms":17113,"temperature":1.0,"reasoning_tokens":2240,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T19:39:52.426919+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct a single particle-number-preserving orbital-rotation context in the (1,1)-sector that reports a measured two-body correlation matrix of rank two or more; this would violate Lemma 2 and break Theorems 4-5. Short of that, measure the Heisenberg witness on the four-qubit code with two orbital-rotation contexts and check whether the residual drops below the predicted Eckart-Young tail.","supporting_citations":[],"review_version":1}