{"id":"0994f16d-313f-46a9-9cb7-730c340529ed","arxiv_id":"2608.04103","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A compact two-qutrit measurement-only protocol on a trapped-ion processor reconstructs the squared braiding phases and fusion amplitudes of D(S3) anyons with average output-state fidelities above 0.998.","lead":"The authors show that non-Abelian braiding and fusion data of the D(S3) anyon model can be extracted on a trapped-ion quantum computer with a compact two-qutrit measurement protocol that never moves anyons. If the result holds, this gives a scalable, low-overhead way to characterise and build topological quantum computing primitives on current hardware.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Unproved cancellation lemma in Appendix A3 (Eq. A22) is the load-bearing step: unless the flux-changing sum vanishes, the measured overlaps are not the D(S3) data claimed.","rationale":"The reader's weakest_assumption identifies exactly the step I consider most load-bearing: Appendix A3's Eq. (A22). I agree. The rest of the protocol is internally coherent: the reduced operators are explicitly given, the fusion and braiding identities in Appendix A4 can be checked by direct multiplication, the experimental counts are reported, and the authors openly state the low acceptance and gauge ambiguities. I do not see an internal inconsistency in the main text's formulas; the Hadamard-test normalization and the post-selected fusion ratios are consistent with the stated dilations. The main reason the verdict should remain CONDITIONAL rather than ACCEPT is precisely the missing proof of the product-state cancellation. It is not grounds for rejection, because the identity is concrete and checkable, and the paper provides enough operator data to verify it. Secondary considerations, such as the absence of propagated uncertainties on F_R and F_F and the use of the model-specific A-B symmetry in Appendix B, are disclosed by the authors and do not shift the verdict. If the proposed exact check passes, I would move to ACCEPT; until then CONDITIONAL is appropriate.","tokens_in":27884,"tokens_out":12033,"duration_ms":113749,"concrete_test":"Run an exact computer-algebra check over the 3x3 qutrit basis: for each operator product O in the set {(F_rho2 F_rho1 A_i F_rho2 F_rho1) for i=+,G; O_m^ji and O_n^j for i,j in {+,G}} with the reduced operators of Eqs. (A12)-(A15), compute (i) Delta(O) = sum_{g1g2 != g3g4} q_{g1g2g3g4} and verify Eq. (A22) is exactly zero; and (ii) verify Eq. (A23), sum_{g in Z3} <Psi_g|O|Psi_g> = 3<xi|O|xi>, using exact rational/cyclotomic arithmetic. A short symbolic script suffices. If all six identities hold, the product-state reduction is exact and the concern is resolved; if any fails, recompute the reported (R_i)^2 and Phi_ij directly from the true four-qudit overlaps to quantify the error in the headline fidelities.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on the chain of reductions in Appendix A, culminating in Eq. (A24)/(31): every four-qudit plaquette overlap <eta|O|eta> in Eqs. (20)-(22) equals (1/4)<xi|O|xi> for the reduced two-qutrit operators. The third reduction (Appendix A3) splits O into flux-preserving and flux-changing matrix elements and asserts, in Eq. (A22), that the total flux-changing contribution cancels 'by direct evaluation'. No derivation, symmetry argument, or verification script is provided. This is not a formality: A+(v) and AG(v) preserve total flux, but the ribbon operators in Eqs. (A12)-(A13) contain Z tensor X, Z^2 tensor X^2, X tensor Z^2, and X^2 tensor Z, each of which changes g1+g2 mod 3, so the cancellation is a nontrivial sum over Z3 matrix elements of products containing up to six such operators. If Eq. (A22) fails for any of the operator products in Eqs. (A1)-(A3), then the extracted p(m_j^i), p(n_j), and Hadamard quadratures are not the four-qudit D(S3) overlaps, and the reconstructed squared braiding phases and fusion amplitudes, together with F_R, F_F, and the magic-resource conclusion, do not follow. The manuscript itself flags this step as omitted, and no machine-checked or independent proof is supplied.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes and experimentally implements a measurement-only protocol for reconstructing the squared braiding phases and squared fusion amplitudes of the D(S3) quantum double, using a reduced two-qutrit encoding on Quantinuum H2 hardware. The authors derive a dense reduction from four-qudit plaquette overlaps to expectation values on a two-qutrit product state, implement the reduced ribbon and projection operators with ancilla-assisted dilations, and extract the braiding data from an adapted Hadamard test and the fusion data from post-selected probabilities. They report average normalized output-state fidelities F_R=0.9988 and F_F=0.9987, and use the reconstructed primitives to construct a non-Clifford braid and a non-stabilizer resource state.","tokens_in":28147,"tokens_out":12304,"duration_ms":111384,"significance":"If the dense reduction is valid, the paper offers a significant step toward practical access to non-Abelian topological data: it replaces extended anyon transport and four-qudit ground-state overlaps with a compact two-qutrit product-state protocol, and it demonstrates that the resulting primitives can be used to produce a non-stabilizer resource. The work has notable strengths: the error bars are propagated from binomial shot noise, the circuits are optimized with state-specific controlled constructions, and the raw data and analysis code are made publicly available. The significance is currently conditional on two load-bearing points: the unproved cancellation in Appendix A3 and the under-reported statistical uncertainty on the F_GG fusion channel.","major_comments":[{"comment":"The product-state replacement is load-bearing and is justified only by the statement that 'direct evaluation shows' that the total flux-changing contribution cancels. The operator products in Eqs. (A1)-(A3) contain ribbon factors F_rho1 and F_rho2 whose individual terms change g1+g2 mod 3, so the claimed cancellation is a nontrivial identity rather than a formality. If Eq. (A22) fails for any of the operator products entering Eqs. (20)-(22), then Eq. (A24) and hence Eq. (31) are incorrect, and the measured overlaps are not the D(S3) braiding and fusion data claimed. Please supply an explicit derivation, a symmetry argument, or a machine-checkable verification for each operator product in Eqs. (A1)-(A3). The same request applies to the assertion in Appendix B, Eq. (B34), that 'direct expansion shows' m_+- = m_-+ = 0; that orthogonality is also used to complete the A-B block of the fusion matrix.","section":"Appendix A 3, Eq. (A22)"},{"comment":"The fusion reconstruction sets c = sqrt(p(m_GG)/p(n_G)) = 0 because zero accepted m_GG events were observed. For N1 = 11,000 and p(n_G) = 4.6e-4, the 95% upper bound on p(m_GG) is approximately 2.7e-4, which gives an upper bound |F_GG| of about 0.77. The data are therefore consistent with the ideal value 0, but they are also consistent with a large non-zero value. Because Eq. (13), Eq. (15) and Eq. (19) use the point estimate c = 0 without propagating its uncertainty, the quoted fusion fidelity and the magic-resource value overstate the constraining power of the experiment. Please report a confidence interval for c and propagate it through F_F and M_2, or explicitly state that c is taken from theory rather than reconstructed from the data.","section":"Section II B and Methods III C 2, p(m_GG) channel"}],"minor_comments":[{"comment":"The notation is confusing: in Eq. (11) the quantities a, b and c are elementwise squared fusion amplitudes, while the entries of F_exp in Eq. (13) are their square roots. Please make this explicit so that the reader does not read the entry 0.497 as (F_AA)^2 rather than F_AA.","section":"Eqs. (11)-(13)"},{"comment":"The statement that both reconstructed phases are separated by more than five standard deviations from zero-amplitude should be quantified with the complex distance and its propagated uncertainty, since the real quadrature of (R_G)^2 is only about 3 standard deviations from zero.","section":"Section II A"},{"comment":"The Monte Carlo propagation used for the uncertainty in M_2 is not described; please provide the number of samples and the method so that the reported value 0.258 ± 0.0886 is reproducible.","section":"Section II C"},{"comment":"The A-B symmetry and the reality of the F-symbol are model inputs used to complete the fusion matrix from the measured +- and G-sector overlaps. Please mark these assumptions clearly in the reconstruction pipeline, for example in a table listing which entries are directly measured and which follow from the imported D(S3) data.","section":"Appendix B and Section II B"},{"comment":"The illustrative sentence 'For 10^4 shots one expects about 5 accepted events' uses a round number, while the actual normalisation run uses 13,000 shots; aligning the numbers would avoid a minor inconsistency.","section":"Section II B"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a plausible fit for the journal and the experimental demonstration is impressive if the reduction is valid. The main risks are the unproved cancellation lemma in Appendix A3 and the under-reported uncertainty on the zero-event F_GG channel; both are fixable within the scope of the manuscript. I do not see grounds for rejection, but the central reconstruction claim should not be accepted until the cancellation is proven and the zero-event channel is quantified with a confidence interval."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a good paper to know about: it shows that the D(S3) braiding and fusion primitives can be pulled off a two-qutrit protocol on H2 without transporting anyons, and the measured phases and fusion amplitudes land close to theory. The new thing is not the reduced operator construction — that is the authors' own earlier work — but the full pipeline: the adapted Hadamard test for non-unitary ribbon overlaps, the post-selected fusion readout, the explicit qubit encodings, and the hardware results. The data are deposited, the shot counts and accepted-event numbers are given, and the error bars are honestly propagated from binomial counting. For a subfield that usually needs 50+ qubits to touch non-Abelian data, this is a genuinely useful modular result.\n\nThe central claims hold up: the measured phases are separated from the depolarized zero by more than five sigma, the fusion matrix is consistent with the ideal F in the real gauge, the c=0 channel is correctly empty, and the magic-resource conclusion follows from the reconstructed parameters without being fit. The average fidelities 0.9988 and 0.9987 are conditional on the accepted branches, and the authors say so; the fusion fidelity in particular rides on a post-selection that keeps 0.046% of the G-normalization trials. That is a real limitation, but it is disclosed and quantified.\n\nThe soft spot is the product-state reduction in Appendix A3. The cancellation of the flux-changing matrix elements is asserted as 'direct evaluation shows' in Eq. (A22), with no derivation, symmetry argument, or script. The stress-test note is right to call this load-bearing: if that sum fails for any of the operator products in Eqs. (A1)–(A3), the measured overlaps are not the four-qudit D(S3) data claimed. I do not think the cancellation is wrong — the operators have enough Z3 structure that it likely holds — but 'likely' is not a proof, and this is exactly the kind of step a referee should make the authors write out. It is a fixable gap, not a crack in the floor.\n\nMy other quibbles are minor: the header fidelities come without propagated uncertainties, and the repeated A–B symmetry is imported from the model rather than measured. Neither changes my read.\n\nWho is this for? People working on topological data extraction, measurement-only anyon protocols, and compact encoded processors. It deserves a serious referee — the experiment is reproducible from the deposited data and the math is checkable — but I would ask for the Appendix A3 proof and a table of the post-selection rates before acceptance.","headline":"A solid, compact hardware demonstration of measurement-only D(S3) data reconstruction, with one unproved algebraic reduction (Eq. A22) that should be pinned down before this is cited as a building block.","tokens_in":28708,"tokens_out":978,"would_cite":true,"duration_ms":11631,"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":"A two-qutrit, measurement-only protocol reconstructs the D(S3) braiding and fusion data without anyon transport, with average output-state fidelities above 0.998, and the combined primitives generate a non-stabilizer resource state.","keywords":["non-Abelian anyons","quantum double D(S3)","measurement-only topological quantum computation","ribbon operators","braiding phases","fusion amplitudes","Hadamard test","non-stabilizer resource states"],"falsifier":"Evaluate the cancellation in Eq. (A22) for each operator product in Eqs. (20)--(22); a single nonzero flux-changing sum would falsify the product-state reduction and hence the reconstructed primitives. A direct numerical computation of the original four-qudit ground-state overlaps would settle the reduction without further hardware.","tokens_in":27649,"feed_emoji":"⚛️","tokens_out":18384,"duration_ms":145040,"temperature":0.7,"pith_summary":"Non-Abelian anyons are usually probed by creating excitations, moving them around each other, and measuring their fusion outcomes—an experimentally heavy process. This paper claims that both essential signatures of the $D(S_3)$ quantum double, the squared braiding phases and the fusion amplitudes, can be reconstructed from temporally ordered ribbon operations and local charge projections acting on just two qutrits, with no physical anyon transport. The claim is backed by a trapped-ion implementation whose reconstructed braiding and fusion transformations reproduce their ideal actions with average normalised output-state fidelities of $\\bar{\\mathcal{F}}_R=0.9988$ and $\\bar{\\mathcal{F}}_F=0.9987$. The recovered primitives also combine into a non-Clifford braid that produces a non-stabilizer resource state, so the measurement-only sequence serves both as a diagnostic of topological data and as a source of magic-state resources.","feed_headline":"Braiding and fusion data recovered without moving anyons","feed_subtitle":"Two-qutrit measurement-only protocol hits ~99.9% fidelity on trapped-ion hardware, then yields a non-stabilizer state.","key_machinery":"The load-bearing mechanism is a dense reduction from the full four-qudit $D(S_3)$ plaquette to two qutrits. Every operator product entering the braiding and fusion overlaps acts nontrivially on only two of the four plaquette qudits; after ordering the group basis by $S_3\\simeq\\mathbb{Z}_3\\rtimes\\mathbb{Z}_2$, the relevant operators have support only on the $\\mathbb{Z}_3$ block, and the flux-changing matrix elements cancel. The overlaps therefore reduce to expectation values on the simple product state $|\\xi\\rangle=\\frac{1}{3}\\sum_{g_1,g_2\\in\\mathbb{Z}_3}|g_1,g_2\\rangle$, with $\\langle\\eta|O|\\eta\\rangle=\\frac{1}{4}\\langle\\xi|O|\\xi\\rangle$. The reduced ribbon operators $F^G_{\\rho_1}=Z\\otimes X+Z^2\\otimes X^2$ and $F^G_{\\rho_2}=X\\otimes Z^2+X^2\\otimes Z$ and the charge projectors $A_+(v)$, $A_G(v)$ carry the non-Abelian action; their non-unitary products are dilated into ancilla-assisted circuits, with an adapted Hadamard test extracting real and imaginary parts of the squared braiding phases and post-selected overlap and normalisation probabilities extracting squared fusion amplitudes. The three-qubit unit-Hamming-weight encoding maps single bit flips outside the logical qutrit subspace, providing code-space filtering.","core_discovery":"This paper establishes that the squared braiding phases and squared fusion amplitudes of the closed fusion subcategory $\\{A,B,G\\}$ of the $D(S_3)$ quantum double can be recovered by a compact two-qutrit protocol. The braiding data are extracted from overlaps of reduced ribbon operators and charge projectors using an adapted Hadamard test: the measured values $(R_+^{\\mathrm{exp}})^2=-0.480\\pm0.095-i(0.853\\pm0.095)$ and $(R_G^{\\mathrm{exp}})^2=-0.418\\pm0.135+i(0.846\\pm0.134)$ agree with $\\bar{\\omega}=-0.5-i0.866$ and $\\omega=-0.5+i0.866$. The fusion data come from post-selected overlap and normalisation probabilities, giving squared amplitudes whose square roots, in the conventional real gauge, give the experimental fusion matrix with entries $0.497\\pm0.022$ and $0.768\\pm0.062$ in place of $1/2$ and $1/\\sqrt{2}$, and with the vanishing $G$--$G$ entry reproduced by an absence of accepted events in the $m_{GG}$ channel. Acting on $10^5$ Haar-random logical qutrit states, the reconstructed transformations achieve average normalised output-state fidelities $\\bar{\\mathcal{F}}_R=0.9988$ and $\\bar{\\mathcal{F}}_F=0.9987$. Combining the primitives into the composite braid $B_s=\\mathbf{F}_s^\\dagger \\mathbf{R}^2 \\mathbf{F}_s$ produces a non-Clifford braid, witnessed by stabilizer Rényi entropy $M_2(|\\psi_B^{\\exp}\\rangle)=0.258\\pm0.0886$ for the output state, compared with the ideal $\\log(16/13)\\approx0.208$, so the data that certify the anyon model also supply a non-stabilizer resource.","pith_inferences":["Editorial inference: the same dense-reduction pattern should transfer to other quantum doubles $D(G)$ whose group splits as a semidirect product and whose ribbon operators act on a fixed small number of qudits, so two-qudit product-state protocols could certify anyonic data for larger models.","Editorial inference: the fusion readout's heavy post-selection, with the $G$-normalisation channel accepting roughly $4.6\\times10^{-4}$ of trials, is the bottleneck for scaling; replacing it with non-destructive syndrome checks or adaptive feed-forward would turn the acceptance overhead into a resource trade-off rather than an exponential sampling cost.","Editorial inference: because the experiment measures only squared amplitudes, the relative sign of fusion channels remains ambiguous; an interference measurement of the phase between two fusion histories would remove the gauge freedom and pin down the full $F$-symbol, not just its magnitudes."],"forward_implications":["The two-qutrit protocol can serve as a modular building block: several such blocks coupled by ribbon and charge-projection operations enlarge the fusion space and enable multi-anyon processes.","The same measurement sequence that reconstructs the topological data also prepares a non-stabilizer resource state, so magic-state generation and data certification can share one circuit.","Characterising D(S3) braiding and fusion no longer requires preparing an extended topological state or physically moving anyons, lowering the overhead for probing non-Abelian anyon models on digital hardware.","The high average output-state fidelities show that the ancilla-assisted dilation and post-selection architecture preserves the relative logical phases needed for encoded anyonic gates across deep circuits."],"supporting_citations":[{"why":"Supplies the measurement-only topological quantum computation principle that braiding can be generated by charge measurements rather than physical transport.","marker":"[4, 5]"},{"why":"Provides the reduced two-qudit D(S3) construction and the non-Clifford braid from which the two-qutrit overlaps are derived.","marker":"[19]"},{"why":"The large-scale trapped-ion D(S3) realisation that the compact two-qutrit protocol replaces as a route to anyonic data.","marker":"[24]"},{"why":"Specifies the trapped-ion processor used to run the implemented circuits.","marker":"[25]"},{"why":"Justifies choosing a real gauge for the F-symbol, so squared-fusion measurements determine the matrix up to sign.","marker":"[26, 34]"},{"why":"Reck decomposition gives the Givens-rotation compilation of the qutrit Fourier transform in the qubit encoding.","marker":"[31]"},{"why":"The Hadamard-test technique adapted to extract real and imaginary parts of the squared braiding phases.","marker":"[33]"}],"fun_headline_variants":["Anyon braiding data without moving anyons","Measurement-only anyon protocol hits 99.9% fidelity","Non-Abelian data from static anyons","Trapped-ion protocol extracts anyon physics without transport"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The protocol assumes that, for every operator product entering the braiding and fusion overlaps, all matrix elements that change the total $\\mathbb{Z}_3$ flux cancel exactly, so the four-qudit ground-state overlap can be evaluated on the simple product state $|\\xi\\rangle$; if any such term survives, the measured overlaps do not equal the intended $D(S_3)$ data.","fun_headline_variants_meta":{"raw":{"variants":["Anyon braiding data without moving anyons","Measurement-only anyon protocol hits 99.9% fidelity","Non-Abelian data from static anyons","Trapped-ion protocol extracts anyon physics without transport"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000162,"raw_usage":{"total_tokens":1359,"prompt_tokens":1184,"completion_tokens":175,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":800,"completion_tokens_details":{"reasoning_tokens":112}},"tokens_in":800,"tokens_out":175,"duration_ms":2327,"temperature":1.0,"reasoning_tokens":112,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:44:06.680301+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the cancellation in Eq. (A22) for each operator product in Eqs. (20)--(22); a single nonzero flux-changing sum would falsify the product-state reduction and hence the reconstructed primitives. A direct numerical computation of the original four-qudit ground-state overlaps would settle the reduction without further hardware.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the reduced two-qudit D(S3) construction and the non-Clifford braid from which the two-qutrit overlaps are derived."},{"cited_title":"& Kitaev, A","cited_arxiv_id":null,"evidence_quote":"The large-scale trapped-ion D(S3) realisation that the compact two-qutrit protocol replaces as a route to anyonic data."}],"review_version":2}