{"id":"16aedfa5-5d1a-4ad9-9082-14323fd7ab78","arxiv_id":"2412.04938","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A unitary decomposition using SWAP and center-switch gates reduces the number of terms needed to encode tridiagonal linear systems in the variational quantum linear solver, with first simulator and hardware demonstrations at 2x2 and 4x4 sizes.","lead":"This paper proposes a new way to split tridiagonal matrices, which appear in simple engineering simulations, into quantum gate pieces for the variational quantum linear solver, reducing the number of pieces by about half. They run examples of size 2x2 and 4x4 on a simulator and on an IBM quantum chip, reaching high solution accuracies despite noisy cost estimates.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The general-n decomposition is asserted rather than proved, and the gate set is under-specified: for n>3 a single 'CS(n−2)' gate is insufficient, and the count depends on an unstated hierarchy of center-switch gates plus 2^{n−1} diagonal Z-correction terms.","rationale":"The reader's weakest-assumption analysis correctly identifies the missing general-n proof and the inconsistent term-count notation as the main risk. My independent reconstruction of the decomposition suggests that the intended construction is actually correct: the off-diagonal edges can be assigned uniquely by the number of trailing ones in the lower index, leading to n unitary off-diagonal gates (X, SWAP, and center-switch gates on qubit blocks 0..k), and the diagonal artifacts of these gates require exactly 2^{n-1} even-parity Z-string corrections. Under this interpretation, Table II's count n + 2^{n-1} is right. However, the paper never states this hierarchy or proves the spanning, and the prose error saying the diagonal correction needs only n unitaries obscures the true count. This is a rigor and reproducibility gap rather than a demonstrated counterexample, so it does not by itself overturn the reader's CONDITIONAL verdict. The proposed checker would settle whether the general claim survives; if it passes, the remaining issue is presentation, not correctness.","tokens_in":11166,"tokens_out":25781,"duration_ms":251854,"concrete_test":"Implement a checker for n=2..8: build the tridiagonal Toeplitz matrix T of size 2^n, generate the candidate unitaries X on q0, SWAP on q0,q1, and center-switch transpositions S_k=(011...1,100...0) on qubits 0..k for k=2..n-1, together with all even-cardinality Z strings on n qubits; solve for coefficients and verify T = sum_l c_l A_l exactly with the number of nonzero coefficients equal to n + 2^{n-1}. If the literal one-gate set containing only the n-qubit CS(n−2) is used instead, the check fails for n≥4, confirming the under-specification.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central efficiency claim is the exact term count 2^{n-1}+n for all n (Section III.C, Table II), but no general proof is given. The text states that the required multi-qubit gates are 'the SWAP and CS(n−2) gates' (Section III.B.2), where CS(n−2) swaps the two n-bit strings 011...1 and 100...0. For n>3, a single such transposition covers only the central edge (2^{n-1}-1, 2^{n-1}); the other interior edges, such as 3-4 and 11-12 for n=4, require center-switch gates on fewer qubits. Table II implicitly uses this hierarchy (I3CS(2−0) and CS(2)(3−0) for n=4), but the general gate family is never stated, and the spanning property is asserted without proof. The diagonal correction is also misdescribed in the text as requiring 'n additional unitaries', whereas Table II and an independent check show the correct count is 2^{n-1} even-parity Z-strings. Thus the claimed saving is real only under a particular unstated interpretation of the gate set; under the literal one-gate reading, the construction is incomplete for n>3.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper applies the variational quantum linear solver (VQLS) to the one-dimensional Poisson problem, whose discrete matrix is tridiagonal and Toeplitz. It proposes a decomposition of such a matrix into Pauli strings together with SWAP and newly introduced 'center-switch' multi-qubit gates, claiming a reduction in the number of unitary terms from 2^n to 2^{n-1}+n for a 2^n x 2^n matrix. The 4x4 and 8x8 cases are worked out explicitly, and simulated plus real-hardware VQLS results are reported for 2x2 and 4x4 systems.","tokens_in":11404,"tokens_out":11167,"duration_ms":101747,"significance":"If the general-n decomposition is correct, the paper offers a concrete, non-trivial instance of a VQLS decomposition for an engineering-relevant matrix class, with a term-count saving that is exponential in n (from 2^n to 2^{n-1}+n). The trade-off against circuit depth is honestly acknowledged. The 4x4 explicit decomposition is correct and the hardware demonstration, though small, addresses a gap in the VQLS literature, which has mostly focused on physics Hamiltonians. However, the central claim for arbitrary n is asserted rather than proved, and the text contains internal inconsistencies about the gate family and the number of diagonal correction terms. The significance is therefore conditional on a rigorous general construction, which the current manuscript does not supply.","major_comments":[{"comment":"The text states that 'The set of Pauli gates and their tensor products, together with the SWAP and CS(n−2) gates are all the unitary terms required to decompose a tridiagonal 2n × 2n matrix' (Section III.B.2, final paragraph). This is not correct for n>3: the single gate CS(n−2) defined in Eq. (20) only swaps the central pair (2^{n-1}-1, 2^{n-1}). For n=4, Table II also needs I3CS(2−0), i.e., a center-switch gate on fewer qubits, covering the pairs (3,4) and (11,12). The general construction requires a hierarchy of center-switch gates on k+1 qubits for k=1,...,n−1 (with k=1 being SWAP and k=0 being X on the least significant qubit). This hierarchy is visible in Table II but never stated in the text. Please define the full gate family explicitly and prove that the n gates cover each of the 2^n−1 tridiagonal off-diagonal edges exactly once.","section":"Section III.B.2, Eq. (20)"},{"comment":"The claim that the diagonal correction 'only requires n additional unitaries' is contradicted by Table II: for n=4 the decomposition contains seven non-identity Z strings (I3I2Z1Z0, I3Z2I1Z0, I3Z2Z1I0, Z3I2I1Z0, Z3I2Z1I0, Z3Z2I1I0, Z3Z2Z1Z0). The correct count of non-identity Z strings is 2^{n-1}-1, so the total term count is n + 1 + (2^{n-1}-1) = 2^{n-1}+n, which matches the stated formula. The explanation in the text must be corrected so that it is consistent with Table II and with the final count.","section":"Section III.C"},{"comment":"No proof is given that the proposed set of unitaries spans the space of tridiagonal matrices of the form (10) for general n, nor are the expansion coefficients specified for general n. The explicit 4x4 and 8x8 examples are not a substitute, because it is not demonstrated that the same pattern of coefficients and Z corrections works for every n. Please provide a constructive proof (for example, by induction, or by explicitly verifying that the n off-diagonal gates cover all edges of the tridiagonal graph and that the even-parity Z strings fix the diagonal) and state the coefficients in Eq. (3) for arbitrary n.","section":"Section III.B.2 and III.C"}],"minor_comments":[{"comment":"Minor language issues: 'less terms' should be 'fewer terms', and 'Proabably' is a typo for 'Probably'.","section":"Abstract and Section I"},{"comment":"The formula in the text '2(n−1) + n' should read '2^{n-1} + n'; the superscript formatting appears to have been lost in compilation.","section":"Section III.C"},{"comment":"The 51/99 circuit-depth comparison in Section III.C is asserted without details of the compiled circuits; for reproducibility, please include the depth calculation method or a code repository.","section":"Section IV"},{"comment":"The hardware results for the 4x4 case consist of only two runs. Please state the number of shots and any error bars, and consider adding more repetitions or at least an explicit caveat about the limited statistics.","section":"Section IV"},{"comment":"Reference [30] is a master's thesis cited as 'to be published'; please provide a stable reference or archive link.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript has the character of an extended abstract: the central general-n claim is not stated as a theorem or proved, and the gate family and term-count explanations are internally inconsistent in places. The underlying construction appears plausible from Table II and the 4x4/8x8 examples, so a careful revision with a rigorous general treatment could make it publishable. The hardware demonstration is modest but acceptable as a proof-of-concept if the technical claims are cleaned up."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear [name],\n\nQuick take: the paper has a genuine new construction, but the general claim is not backed up. The 4x4 and 8x8 decompositions are shown explicitly and they work; the term count for those cases is right. The paper also reports first VQLS runs on tridiagonal systems, including real hardware. That part is fine.\n\nThe soft spot is the step from examples to the general formula. The text says the multi-qubit gates are SWAP and CS(n−2), but a single CS(n−2) only swaps the two central bitstrings. For n>3, the other interior edges require center-switch gates on fewer qubits. Table II implicitly uses such a hierarchy, but the text never defines it. The diagonal correction is also misdescribed: the text says n unitaries, while the table (and direct computation) gives 2^{n-1} Z strings. So the stated count 2^{n-1}+n is not actually derived. It might be right under some interpretation, but the paper as written does not pin that down.\n\nThe experiments are preliminary: two hardware runs for 4x4, no code, no error bars. That is acceptable for a note, but not enough to justify the efficiency claim. And since tridiagonal systems are classically trivial (Thomas algorithm), the practical relevance is modest.\n\nStill, the construction is concrete and checkable. A referee can verify the small examples and push the authors to provide a rigorous general derivation and consistent notation. I'd send it to review, but I'd expect major revision.","headline":"A concrete small-case decomposition for tridiagonal VQLS, but the general-n claim is asserted, not proved.","tokens_in":11975,"tokens_out":6639,"would_cite":false,"duration_ms":57728,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"This paper shows that tridiagonal matrices from 1D Poisson problems can be decomposed into $2^{n-1}+n$ unitary terms using SWAP and center-switch gates, enabling the first VQLS runs on such systems.","keywords":["unitary decomposition","variational quantum linear solver","tridiagonal systems","Poisson problem","center-switch gate","Pauli strings","NISQ devices","quantum linear system problem"],"falsifier":"For a $16\\times 16$ tridiagonal matrix ($n=4$), explicitly construct the paper's claimed 12 unitary terms, sum them, and compare against the original matrix; any missing or extra entry would refute the counting claim, and the same check can be repeated for $n=5$ with the claimed 21 terms.","tokens_in":10945,"feed_emoji":"⚛️","tokens_out":6795,"duration_ms":59267,"temperature":0.7,"pith_summary":"Variational quantum linear solvers (VQLS) have previously been demonstrated on matrices that are naturally sums of a few unitary terms, such as small-qubit Hamiltonians. This paper targets tridiagonal matrices with constant coefficients, which arise from discretizing the one-dimensional Poisson problem, and shows how to decompose a $2^n\\times 2^n$ matrix of this type into $2^{n-1}+n$ unitary terms using Pauli strings together with SWAP and center-switch gates, instead of the $2^n$ Pauli strings required by the standard decomposition. Because VQLS evaluates $O(L^2)$ quantum circuits when the matrix has $L$ terms, cutting $L$ almost in half substantially reduces the cost-circuit count. The paper reports the first simulated and real-hardware VQLS solutions of such tridiagonal systems, with solution fidelities close to 1 even when the cost function estimate is noisy. The tradeoff is that the multi-qubit gates make the cost circuits deeper, illustrated by a $51/99$ maximum-depth ratio for a $4\\times4$ example.","feed_headline":"New gate set halves terms for quantum tridiagonal solver","feed_subtitle":"Fewer unitary terms let a variational quantum linear solver tackle 1D Poisson problems for the first time.","key_machinery":"The central object is the center-switch gate $\\mathrm{CS}(n-2)$, a permutation matrix on $n$ qubits that swaps the computational basis states $|011\\ldots1\\rangle$ and $|100\\ldots0\\rangle$; for $n=2$ it reduces to the standard SWAP gate. The paper decomposes it as a sequence of $2n-1$ multi-controlled Toffoli gates and uses it, together with Pauli strings and $Z$-type tensor products, to reproduce the tridiagonal structure of the matrix $A$ in the VQLS cost function, reducing the number of terms in the expansion from $2^n$ to $2^{n-1}+n$.","core_discovery":"The central claim is that the class of tridiagonal Toeplitz matrices used in 1D Poisson problems admits a unitary decomposition with roughly half as many terms as the standard Pauli-string expansion, by adding multi-qubit permutation gates. The key observation is that the center off-diagonal entries of a tridiagonal matrix are exactly the off-diagonal entries of a SWAP gate for the $4\\times4$ case, and of a generalized center-switch gate $\\mathrm{CS}(n-2)$ that swaps the bitstrings $011\\ldots1$ and $100\\ldots0$ for larger $n$. The paper constructs these gates explicitly in terms of Toffoli gates and compensates their diagonal entries with $n$ tensor products of $Z$ operators, leading to the count $2^{n-1}+n$. It then uses this decomposition inside VQLS to solve $2\\times2$ and $4\\times4$ linear systems, obtaining fidelities close to 1 in simulation and on hardware, and documents the tradeoff between fewer terms and deeper quantum circuits.","pith_inferences":["The $O(L^2)$ circuit count means the term reduction from $2^n$ to about $2^{n-1}$ should cut the number of cost-circuit evaluations by roughly a factor of four, a consequence the paper does not state explicitly.","The same permutation-gate construction could be adapted to other banded or structured matrices that have off-diagonal blocks resembling SWAP-like permutations, though the diagonal compensation would need to be re-derived.","For larger $n$, the center-switch gate requires $2n-1$ multi-controlled Toffoli gates, so one could test whether alternative decompositions of the permutation yield shallower circuits and whether the depth penalty stays acceptable on current hardware.","If the term count holds at scale, the main obstacle to using this decomposition becomes circuit depth rather than circuit count, shifting the focus to depth-reduction techniques such as the Hadamard-Overlap Test."],"forward_implications":["For a $2^n\\times 2^n$ tridiagonal system, the number of unitary terms in the matrix expansion drops from $2^n$ to $2^{n-1}+n$, reducing the number of quantum circuits needed to estimate the VQLS cost function.","The multi-qubit-gate decomposition extends VQLS beyond few-qubit Hamiltonians to a class of engineering matrices derived from finite-element discretizations.","The depth penalty is real but not prohibitive at small scale: the $4\\times4$ case runs in circuits of depth 99 versus 51, and still reaches high fidelity on hardware.","The small hardware runs show that even when the cost estimate is biased by sampling noise and decoherence, the optimizer can land at parameters close to the analytical solution, a resilience the authors attribute to noise tolerance of variational algorithms."],"supporting_citations":[{"why":"Supplies the VQLS algorithm, cost function, and Hadamard-Overlap Test that the new decomposition is designed to feed.","marker":"[14]"},{"why":"Identifies tridiagonal systems as the discretization of the 1D Poisson problem, the target application.","marker":"[18]"},{"why":"Provides the general Pauli-string decomposition into $N^2$ terms that constitutes the baseline the paper improves on.","marker":"[29]"},{"why":"Cited for the fact that tridiagonal matrices of the considered form have only $N$ non-zero Pauli coefficients, the starting point of the comparison.","marker":"[30]"}],"fun_headline_variants":["Multi-qubit gates trim quantum linear solver terms","VQLS tackles Poisson with half the unitary terms","New decomposition cuts quantum tridiagonal cost","Quantum solver: fewer terms, deeper circuits","Half the gates: VQLS solves tridiagonal systems"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central efficiency claim depends on the unproved assumption that the Pauli strings, SWAP gate, and center-switch gates $\\mathrm{CS}(n-2)$ span the space of tridiagonal Toeplitz matrices for every $n$ and that the term count $2^{n-1}+n$ is exact; if the count is off for larger $n$, the advantage over the Pauli decomposition disappears.","fun_headline_variants_meta":{"raw":{"variants":["Multi-qubit gates trim quantum linear solver terms","VQLS tackles Poisson with half the unitary terms","New decomposition cuts quantum tridiagonal cost","Quantum solver: fewer terms, deeper circuits","Half the gates: VQLS solves tridiagonal systems"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000227,"raw_usage":{"total_tokens":1523,"prompt_tokens":1047,"completion_tokens":476,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":663,"completion_tokens_details":{"reasoning_tokens":413}},"tokens_in":663,"tokens_out":476,"duration_ms":5312,"temperature":1.0,"reasoning_tokens":413,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T21:07:33.740364+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"For a $16\\times 16$ tridiagonal matrix ($n=4$), explicitly construct the paper's claimed 12 unitary terms, sum them, and compare against the original matrix; any missing or extra entry would refute the counting claim, and the same check can be repeated for $n=5$ with the claimed 21 terms.","supporting_citations":[{"cited_title":"Quarteroni, Numerical Models for Differential Problems","cited_arxiv_id":null,"evidence_quote":"Identifies tridiagonal systems as the discretization of the 1D Poisson problem, the target application."},{"cited_title":"A variational linear solver for the poisson 1d matrix,","cited_arxiv_id":null,"evidence_quote":"Cited for the fact that tridiagonal matrices of the considered form have only $N$ non-zero Pauli coefficients, the starting point of the comparison."}],"review_version":1}