{"id":"d580345c-4153-4757-9891-49402249c5b3","arxiv_id":"2607.05650","paper_version":1,"verdict":"ACCEPT","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"Among information-theoretic MBQC VBQC clients, measurement-based RSP and cavity-reflection emission clients are the strongest near-term defaults once noise-robust security, rate, errors, and hardware cost are weighed together.","lead":"This paper compares client device designs for verifiable blind quantum computing with a matter-qubit server, scoring them on security proofs, rate, errors, and hardware cost. It gives experimental groups a decision framework and flags measurement-based remote state preparation and cavity-reflection clients as strong defaults.","discovery_kind":"extension","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified","rationale":"The reader correctly identifies that the ranking of defaults rests on the current literature status of noise robustness. That status is an explicit, openly acknowledged premise of the paper, not a concealed assumption. Because the work is a comparative design analysis whose strongest claim is the framework itself (and only secondarily the provisional defaults), the premise does not create a load-bearing correctness risk. The derived rate and cutoff formulas stand independently of which security proofs are later filled in. Therefore no adjustment to the ACCEPT verdict is warranted.","tokens_in":37624,"tokens_out":342,"duration_ms":50500,"concrete_test":"Check whether any peer-reviewed noise-robust composable verification scheme for receive-and-measure or fixed-gate clients has appeared since the arXiv date; if none exists, the paper’s near-term filter remains valid and the default-candidate statement is unchanged.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper’s central claim is a multi-axis comparative framework, not a hard ranking. Measurement-based RSP and cavity-reflection clients are presented as strong defaults only under the explicit, transparent filter of existing noise-robust prepare-and-send proofs (§IIIA, Table I, §IV). The authors repeatedly state that the choice remains context-dependent and that missing proofs for direct-measurement/fixed-gate clients may change the picture. Rate equations (1–11), cutoff analysis (Appendix A, Eqs. A19, A51, A63), and error inventory (Table III) are derived consistently with the stated scope. No internal contradiction or hidden assumption undermines the framework itself.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript provides a comparative analysis of client architectures for single-server, single-client, information-theoretically secure verifiable blind quantum computing (VBQC) based on measurement-based quantum computation with a matter-qubit server. It organizes existing proposals into emission-based, measurement-based, and rotation-based categories (with prepare-and-send vs receive-and-measure roles), then evaluates them along four axes: existing security proofs and overheads (Table I), derived rate expressions including loss, attempt duration, and cutoff-limited graph generation (Table II, Eqs. 1–20, Appendix A), dominant error modes (Table III), and hardware cost/complexity. Under the stated scope and the practical filter of noise-robust prepare-and-send verification, measurement-based remote state preparation and single-photon cavity-reflection clients are identified as strong default candidates, while the authors emphasize that the optimal choice remains context-dependent and supply a multi-axis decision framework rather than a hard ranking.","tokens_in":37813,"tokens_out":1260,"duration_ms":21607,"significance":"This is a timely systems-level contribution for experimental and architectural work on delegated quantum computing. The field has accumulated many client proposals with incompatible assumptions; a carefully scoped taxonomy that ties each architecture to concrete security proofs, rate scalings, and error inventories is useful for near-term hardware choices. Strengths include: (i) explicit, non-overclaimed security mapping with caveats for non-composable or incomplete proofs (Table I); (ii) architecture-specific success probabilities and attempt times (Eqs. 1–20); (iii) a rigorous renewal-process treatment of measure-as-you-go graph generation under qubitwise cutoffs, with closed forms for linear graphs and columnwise bounds for brickwork graphs (Appendix A, Eqs. A19, A51, A63); and (iv) transparent framing that the ranking is literature-status-dependent and context-dependent. The work does not invent free parameters to force a ranking; free parameters (c, p, μ, k, L, α) are physical or protocol inputs. If the comparative framework is adopted, it should help experimental groups prioritize client designs without re-deriving the security and rate landscape from scratch.","major_comments":[{"comment":"Section IIIB and Table II derive architecture-specific p and t, and Appendix A gives E(K) under cutoffs, but the manuscript never evaluates end-to-end expected round time under a single shared, realistic parameter set (e.g., fixed L, α, p_emit, p_detect, C_i, t_emit, t_switch). The qualitative ranking in §IV (measurement RSP and cavity reflection as strong defaults) therefore rests on structural comparisons (loss exponents, HOM matching, compounding round-trips) plus partial illustrations (transmission-only R_X at L=25 km; brickwork bounds in Figs. 3–4). A compact numerical comparison—even for one metropolitan distance and a few efficiency points—would make the rate axis load-bearing rather than largely structural, and would clarify when rotation-based clients remain competitive at short range.","section":null},{"comment":"Section IIIA and §IV treat absence of noise-robust verification as a near-term disqualifier for direct-measurement and fixed-gate clients. That filter is stated transparently and is consistent with the cited literature (e.g., Takeuchi–Morimae on honest noise), but it is load-bearing for the ‘strong default’ claim. The manuscript should state more explicitly what would reverse the ranking (e.g., a composable noise-robust receive-and-measure proof, or a fixed-gate reduction into prepare-and-send with one-time pads), so that the framework remains usable if those proofs appear. This is a framing fix, not a re-derivation.","section":null}],"minor_comments":[{"comment":"Typo: ‘Leicthel et al.’ should be ‘Leichtle et al.’ (Section IIIA, security overhead discussion).","section":null},{"comment":"Capitalization of ‘Hayashi-morimae’ is inconsistent; use ‘Hayashi–Morimae’ throughout.","section":null},{"comment":"Table I legend uses filled/open/blue-grey circles; ensure the rendered symbols remain distinguishable in grayscale print and that the ‘caveat’ cases (blue-grey) are cross-referenced to the exact paragraph that explains each caveat.","section":null},{"comment":"Equation (3) for the multi-photon factor g is dense; a short sentence stating the regime in which g→1 (μ→0) and when g>1 improves rate at fidelity cost would help non-specialists.","section":null},{"comment":"Figure 2 caption and surrounding text discuss mid-point vs end-point heralding; consider adding the corresponding t expressions (Eqs. 12–15) to the figure for quick reference.","section":null},{"comment":"In Appendix A, the Mathematica notebook link is useful; please confirm the repository will remain available and that the notebook reproduces Figs. 3–4 from the stated parameters.","section":null},{"comment":"The FeMoco estimate (~46 h absolute lower bound at L=25 km) is effective; state explicitly that this ignores security overhead and p≪n_r/c resets so readers do not treat it as a protocol runtime.","section":null}],"recommendation":"minor_revision","confidential_remarks":"This is a solid comparative/systems paper rather than a theorem paper. Fit for a quant-ph journal that publishes architecture and experimental-design analyses is good. No novelty or citation-pattern concerns stood out; the security caveats are handled more carefully than is common in this subfield. I would not require new security proofs as a condition of acceptance."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a practical decision paper for labs building matter-qubit VBQC clients, not a foundational result. What is new is the side-by-side taxonomy (emission / measurement / rotation, with BSM vs cavity reflection variants) plus the derived success probabilities, attempt times, and especially Appendix A’s renewal-process analysis of expected attempts for linear and brickwork graphs under qubitwise cutoffs. Those formulas (and the columnwise bounds) are the part people will actually use.\n\nThey do the security mapping carefully: Table I tracks which architectures inherit noise-robust prepare-and-send proofs, which only have inverse-polynomial or incomplete claims, and what the overhead looks like (rounds vs traps/dummies vs WCP gadgets). Rate and error inventories (Tables II–III) are consistent with standard loss, communication-time, and HOM/cavity models; free parameters (p, c, μ, L, α) are external, not fitted to force a ranking. The abstract and §IV repeatedly say the choice is context-dependent; measurement-RSP and cavity-reflection are “strong defaults” only under the transparent filter of existing noise-robust proofs and fewer compounding error modes. That is honest scoping, not a hard ranking dressed up as one.\n\nSoft spots are real but proportional. The ranking leans on the current literature status of noise robustness; if someone soon supplies a noise-robust composable proof for direct measurement or fixed-gate clients, the “not near-term ready” filter softens. The multi-pass rotation clients look bad on rate for the usual compounding reasons; that is physics, not a modeling error. No code or numerical notebook is shipped in the text (they point to a GitLab repo for the brickwork bounds), and there is no experimental validation—this is pure comparative theory. Citation pattern is appropriate; they cite the protocols they reduce to.\n\nWho it is for: experimental groups choosing client hardware and theorists who need end-to-end rate estimates under cutoffs. It deserves a serious referee. I would bring it to reading group if we are working on networked or delegated QC, and I would cite the rate/cutoff material. Accept for peer review.","headline":"Solid multi-axis comparison of VBQC client hardware with usable rate/cutoff formulas; defaults are carefully scoped, not oversold.","tokens_in":38412,"tokens_out":522,"would_cite":true,"duration_ms":11184,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"Two client designs stand out for verifiable blind quantum computing: measurement-based remote state preparation and cavity reflection.","keywords":["verifiable blind quantum computing","client architectures","measurement-based quantum computation","remote state preparation","cavity reflection","noise robustness","graph-state cutoffs","matter-qubit server"],"falsifier":"Build or simulate a complete end-to-end protocol for a brickwork graph of realistic size under laboratory loss, coherence, and dark-count numbers; if a direct-measurement or multi-pass fixed-gate client finishes with higher verified success rate and lower abort probability than measurement-RSP or cavity-reflection under the same noise, the default ranking is overturned.","tokens_in":38532,"feed_emoji":"🔐","tokens_out":757,"duration_ms":6691,"temperature":0.7,"pith_summary":"Building a quantum computer is expensive, so a thin client device that can safely hand a computation to a remote server without leaking the input, the program, or the answer is valuable. This paper restricts attention to single-server, information-theoretic protocols that use measurement-based quantum computation and a matter-qubit server, then systematically compares every client architecture that fits that scope. Clients are grouped into emission-based, measurement-based, and rotation-based families, with sub-variants that differ in how photons are prepared, measured, or rotated and how they couple back into the server. For each design the authors inventory existing security proofs (composability, noise robustness, generality, overhead), derive attempt success probabilities and expected graph-generation times under memory cutoffs, catalogue dominant error modes, and weigh hardware cost and server requirements. Measurement-based remote state preparation and single-photon cavity-reflection teleportation emerge as the strongest default near-term candidates; they combine access to noise-robust prepare-and-send proofs, relatively mild loss scaling, fewer compounding errors, and no dual-source photon-matching demand. The paper therefore supplies both a ranking and a decision framework so that labs can choose according to the constraints they actually face.","feed_headline":"Two client designs lead for secure delegated quantum computing","feed_subtitle":"Measurement-RSP and cavity reflection win on security, rate, and error modes; a framework picks the rest","key_machinery":"A three-family taxonomy (emission-based, measurement-based, rotation-based) together with explicit rate formulas (success probability per attempt, time per attempt, expected attempts under qubit-wise or column-wise cutoffs for linear and brickwork graphs) and side-by-side tables of security guarantees, error modes, and hardware requirements that convert qualitative architectural differences into quantitative comparison axes.","core_discovery":"Within the stated scope, client architectures that implement measurement-based remote state preparation or single-photon cavity-reflection interaction are the strongest default candidates for near-term verifiable blind quantum computing, because they simultaneously enjoy noise-robust prepare-and-send security proofs, favorable loss structure, limited compounding error modes, and no dual-source Hong-Ou-Mandel matching requirement; the final choice remains setting-dependent and is guided by the multi-axis comparison the authors provide.","pith_inferences":[],"forward_implications":[],"fun_headline_variants":["Measurement-RSP and cavity reflection lead VBQC client designs","Two client classes top defaults for verifiable blind QC","Prepare-and-send proofs favor measurement and reflection clients","Framework guides VBQC client choice beyond two strong defaults","Emission, measurement, rotation clients compared for matter servers"],"cache_read_input_tokens":19712,"weakest_assumption_plain":"The ranking treats the present absence of noise-robust verification proofs for direct-measurement and fixed-gate clients as a hard near-term filter; if those proofs appear soon, or if honest noise proves milder than assumed, the preference ordering can change.","fun_headline_variants_meta":{"raw":{"variants":["Measurement-RSP and cavity reflection lead VBQC client designs","Two client classes top defaults for verifiable blind QC","Prepare-and-send proofs favor measurement and reflection clients","Framework guides VBQC client choice beyond two strong defaults","Emission, measurement, rotation clients compared for matter servers"]},"model":"grok-4.5","effort":"low","cost_usd":0.002528,"raw_usage":{"total_tokens":982,"prompt_tokens":841,"num_sources_used":0,"completion_tokens":80,"cost_in_usd_ticks":25280000,"prompt_tokens_details":{"text_tokens":841,"audio_tokens":0,"image_tokens":0,"cached_tokens":0},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":61,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":841,"tokens_out":80,"duration_ms":1457,"temperature":1.0,"reasoning_tokens":61,"cache_read_input_tokens":0,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-11T04:19:25.025206+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"Build or simulate a complete end-to-end protocol for a brickwork graph of realistic size under laboratory loss, coherence, and dark-count numbers; if a direct-measurement or multi-pass fixed-gate client finishes with higher verified success rate and lower abort probability than measurement-RSP or cavity-reflection under the same noise, the default ranking is overturned.","supporting_citations":[],"review_version":1}