{"id":"7173db0a-319a-42ae-831e-535ad71facd3","arxiv_id":"2412.13151","paper_version":1,"verdict":"UNVERDICTED","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proposes a Model Cards framework adapted from AI to document use cases, performance, and limitations of quantum technology entities.","lead":"This paper proposes a standardized documentation format, Model Cards, for reporting the capabilities and limitations of quantum technologies. It aims to help buyers, engineers, and regulators compare quantum computers, communication systems, and sensors.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Template completeness is asserted, not demonstrated: no worked example is provided, and the paper's own admitted omission of data provenance exposes the main gap for a documentation standard.","rationale":"The paper is a clearly written position piece. Its strongest claim is modest: it proposes a documentation framework. For such a claim, the most load-bearing condition is that the framework is complete enough to be used. That condition is unsupported because no example card is constructed. The template includes fields that are conceptually demanding ('Non-local Quantum Coherence', 'Fundamental limit', 'Assurability' evidence) and whose practical definition is not explored. The paper self-identifies missing data provenance in §IV and relies on voluntary transparency for usefulness. These are not internal contradictions, but they are unverified assumptions about the feasibility of the proposal. My proposed test—a comparative carding exercise on a real device—would directly probe whether the template can be filled consistently. If it can, the proposal gains empirical grounding; if not, the 'generalised and extensible framework' claim is weakened. The reader's weakest assumption overlapped with this (sufficiency/generality without case study); I did not find the voluntary-transparency concern to be the primary gap because any voluntary standard shares it. Therefore the appropriate verdict remains UNVERDICTED, unchanged from the reader.","tokens_in":769,"tokens_out":716,"duration_ms":38227,"concrete_test":"Pick one publicly documented quantum device (e.g., a transmon qubit module with published T1/T2, gate fidelities, and wiring diagram). Ask two independent teams to fill a Model Card strictly using the §III.A–I template and only public information. Then check: (a) can every required field be filled non-vacuously without new experiments; (b) do the two cards agree on intended use, limitations, and metrics; (c) is the 'Fundamental limit' entry derived or cited. If a large fraction of fields is empty, contradictory, or dependent on unstated measurements, the template's sufficiency and generality claim fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"For the proposal to achieve its stated aims of transparency, technology selection, and risk management, a Model Card must be completable for real quantum entities without vacuous or invented entries. The paper specifies fields in §III.A–I but never applies the template to a single concrete device, so there is no evidence that all fields are well-defined. Sections such as 'Non-local Quantum Coherence' (§III.D.2), 'Impact Estimates' (§III.E), and 'Fundamental limit' (§III.F) presuppose that meaningful values exist and can be obtained; for many sensing or communication systems this is far from obvious. The concluding remarks (§IV) concede a 'strong omission' of data provenance, which undercuts the FAIR-compliance recommendation made in §III. Furthermore, the stated usefulness condition depends on actors voluntarily supplying accurate data (§IV), and no mechanism or motivation is proposed. Because the central claim is a normative proposal rather than a tested standard, the lack of a worked example is a genuine load-bearing gap: it leaves the completeness and internal consistency of the template unestablished.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript proposes extending the Model Cards documentation framework, originally developed for machine learning models in ref. [2], to quantum technologies. The authors argue that emerging quantum technologies in computing, communication, and sensing need standardized, transparent documentation covering intended use, performance characteristics, evaluation conditions, and assurance information. The paper presents a detailed template with ten sections: Entity Details, Intended Use, Factors, Quantum Technology Specifications, Errors, Performance Metrics, Ethical Considerations, Evaluation Criteria, Assurability, and Supplementary Materials. The proposal additionally recommends FAIR-compliant metadata and emphasizes extensibility to accommodate future standards. The manuscript is explicitly positioned as a proposal to stimulate community discussion rather than a validated or finalized standard.","tokens_in":7862,"tokens_out":7137,"duration_ms":62117,"significance":"Strengths: the proposal is internally consistent, well-referenced, and builds on an established concept (AI Model Cards) with known traction for transparency reporting; it adds quantum-specific content (coherence, entanglement, QBER, FMEA, assurance cases, FAIR metadata) that goes beyond naive adaptation; and it is honest about its own limitations, explicitly flagging the data-provenance gap in Section IV. Weaknesses: the paper provides no worked example, no data, and no case study, so the claim that the template is complete and general enough for all quantum technology types is asserted rather than demonstrated; and the stated usefulness condition in Section IV depends on voluntary, accurate participation with no proposed mechanism to secure it. If validated by application to real entities, the framework could genuinely support technology selection, risk management, and regulatory assurance for quantum technologies; as it stands, its value is programmatic rather than demonstrated.","major_comments":[{"comment":"The central claim that the proposed template is complete, general, and extensible for all quantum technologies is never demonstrated. The template is not applied to a single concrete entity——not a quantum computer, a QKD system, or a quantum sensor——so there is no evidence that every field is well-defined and completable with meaningful, non-vacuous entries. This is load-bearing because the paper's purpose is to propose a documentation standard for stakeholders to adopt; a field that cannot be filled for real entities would undermine the uses claimed in the abstract (technology selection, risk management, compliance). The concluding remarks defer validation to future stakeholder feedback (Section IV), confirming that the completeness claim is currently unsubstantiated. I recommend adding at least one fully worked example, together with a per-section analysis of which fields are mandatory, conditional, or not applicable for each of the three main technology families (computing, communication, sensing), or explicitly re-scoping the claim as a draft template pending such a study.","section":"Section III.A–I and Section IV"},{"comment":"The paper recommends that model cards conform to the FAIR Guiding Principles (ref. [5]), stating 'we make a further recommendation beyond that of [2] and assert that model cards be created using appropriate metadata conforming to FAIR Guiding Principles.' However, Section IV concedes that data provenance is 'one strong omission in our current proposal.' Provenance is not an optional refinement: under FAIR, reuse requires rich provenance metadata (R1.2), and traceability of measurement conditions is arguably central to a documentation standard whose stated aims include transparency and risk management. As written, the FAIR recommendation and the admitted omission are in direct tension: a card whose metrics cannot be traced to their measurement conditions cannot be FAIR-compliant, nor can it support the assurance claims made in Section I. The authors should either integrate a provenance mechanism (for example, a required measurement-conditions and data-lineage block in Sections E and F, or explicit linkage to a provenance standard) or explicitly delimit the FAIR recommendation.","section":"Section III (FAIR recommendation) vs. Section IV"},{"comment":"Several fields presuppose that well-defined, obtainable quantitative values exist for entities of every type. 'Non-local Quantum Coherence' (III.D.2) requires an enumeration of all entanglement resources; 'Impact Estimates' (III.E) requires quantitative or qualitative impact for each error source; and 'Fundamental limit' (III.F) requires a 'theoretical best possible performance' for each metric. For many real systems, particularly sensing and communication devices and noisy intermediate-scale processors, such values may be unknown, contested, or not well-defined at the level the template implies. The paper asserts that these fields should be filled but gives no account of how to determine them, and no example. Without a demonstrated method for populating these fields, the template risks producing empty or invented entries, which would defeat the transparency goal. A worked example or an explicit field-level feasibility analysis is needed.","section":"Sections III.D.2, III.E, III.F"},{"comment":"The manuscript's own success condition is behavioral: 'The usefulness and accuracy of this approach will depend on those engaging with it to act with transparency and include sufficient data on assurability, verification, and validation.' The paper proposes no mechanism, incentive, or independent validation step to secure accurate self-reporting, yet the abstract claims the approach will enable stakeholders to 'manage risk and assure compliance with regulatory frameworks.' For a documentation standard aimed at risk management and regulatory assurance, reliance on unverified self-reporting is a load-bearing concern. This is not by itself a reason to reject a proposal paper, but it should be addressed in the text——for example, by discussing what role independent audits, standardized benchmarks, or a registration body could play, or by referencing how the AI Model Cards and Datasheets literature has handled the analogous problem.","section":"Section IV (Concluding Remarks)"}],"minor_comments":[{"comment":"The phrase 'unital proposal' appears to be a typo for 'initial proposal'; if the word is intentional, it should be clarified.","section":"Section I"},{"comment":"The unresolved citation placeholder '[ ? ]' after 'as either the ITRS or the IDRS have benefitted from [ ? ]' must be replaced with a reference or removed.","section":"Section II"},{"comment":"The affiliation 'Loughborough Universi ty' contains an extra space between 'Universi' and 'ty'.","section":"Author affiliations"},{"comment":"In the Purpose example, 'a magnetic flux sensor for non-destructive testing in of defects in aircraft composite structures' contains 'in of'; the intended phrasing is presumably 'testing of defects' or 'testing for defects'.","section":"Section III.A"},{"comment":"The text 'For composite carries such as logical qubits' should read 'carriers', and the phrase 'with ζ having configurable φ and r within the region of 4' is unclear about the intended parameter range and units.","section":"Section III.D.2"},{"comment":"The sentence 'Information covered by the above points is inherently connected through casual chains' should read 'causal chains'.","section":"Section III.E"},{"comment":"The sentence 'The next is to this reporting structure based on feedback obtained from a range of stakeholders' is garbled; the intended meaning appears to be 'The next step is to refine this reporting structure based on feedback.'","section":"Section IV"},{"comment":"The URL for reference [8] is truncated in the bibliography ('https://incose.onlinelibrary.wiley.com/doi/pdf/10.1002/j.2334-5837.20'); a complete DOI or stable URL should be provided.","section":"References"},{"comment":"The paper would benefit from engaging with known critiques of the original AI Model Cards framework as experienced in the AI community, such as the burden of completing many fields and the difficulty of verifying self-reported information, since these experiences directly inform design choices in Sections F and I.","section":"Section II and Section III"}],"recommendation":"major_revision","confidential_remarks":"This is a position paper with no original data, derivation, or case study; its contribution is a proposed template. Whether it fits a research journal in quantum physics depends on whether the venue publishes programmatic proposals; a forum/commentary section or a standards-oriented venue might be a better fit. The missing worked example is, in my view, fixable within the manuscript's scope and should be required before acceptance. The data-provenance gap that the authors themselves acknowledge in Section IV should also be closed rather than deferred, since it directly affects the credibility of the FAIR recommendation. I see no circularity concerns and no citation-related red flags."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a reasonable position paper adapting AI Model Cards to quantum technologies. The template is detailed and the writing is honest about its own gaps, but the absence of a single worked example leaves the central claim of completeness and usability unproven. Worth sending to peer review as a position piece, expecting the authors to test the template on a real device.\n\nWhat's new: the quantum-specific sections—coherence, entanglement, error budgets, assurability—go well beyond Mitchell et al., and the FAIR metadata recommendation is a sensible addition. The paper correctly identifies that quantum systems have failure modes and assurance needs that differ from classical ML. The focus on technology selection, risk management, and policy is a useful framing that the AI model-card literature has only partially addressed.\n\nWhat it does well: the structure is clear, the template is laid out in enough detail to be actionable, and the paper explicitly acknowledges its main omission (data provenance) in the concluding remarks. It also resists overclaiming: it calls itself a proposal to stimulate discussion, and notes that a final form needs community agreement. Citations are appropriate, with the AI model card and datasheets papers correctly identified as the basis.\n\nSoft spots: the stress-test critique is on target. No worked example means we don't know whether fields like 'Non-local Quantum Coherence' (Section III.D.2), 'Impact Estimates' (III.E), or 'Fundamental limit' (III.F) have well-defined values for a real magnetometer, QKD link, or annealing processor. The admitted data-provenance gap is real and does undercut the FAIR-compliance recommendation. And the paper's own caveat about voluntary transparency is a structural issue: model cards only work if stakeholders fill them in accurately, and no incentive mechanism is proposed. These are not fatal for a position paper, but they are the difference between a proposal and a standard.\n\nI'd treat the 'UNVERDICTED' label as fine — this isn't a testable research claim. But if I were an editor, I'd still send it out: the quantum-standards community needs proposals from which to argue. My recommendation: accept for peer review, but the referee should ask for at least one worked example and a more concrete handling of provenance. It's a useful contribution for people working on benchmarking, tech evaluation, and roadmapping.","headline":"A decent position paper adapting AI Model Cards to quantum technologies; the template is thoughtful but untested, and the missing worked example is its main weakness.","tokens_in":8370,"tokens_out":3102,"would_cite":false,"duration_ms":27745,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["03.67.-a"],"model":"deepseek-v4-flash","headline":"The paper argues that every quantum technology should be documented with a standardized Model Card stating intended uses, quantum specifications, errors, performance metrics, and assurance evidence.","keywords":["model cards","quantum technologies","documentation standard","transparency","assurability","performance metrics","technology selection","quantum benchmarking"],"falsifier":"Take any existing quantum device—say, a quantum key distribution link or a small superconducting processor—fill in the proposed card from the developer's documentation, and have a second team independently reproduce the performance metrics by following the card's measurement and evaluation sections. If the card cannot be completed without recourse to undocumented assumptions, or if independent replication yields materially different numbers, the template fails as a transparency device. A simpler observational test: if model cards see widespread use and yet almost none report residual errors, limitations, or failed evaluations, the voluntary-honesty premise has not held.","tokens_in":7506,"feed_emoji":"⚛️","tokens_out":7196,"duration_ms":66219,"temperature":0.7,"pith_summary":"This paper proposes that quantum technologies—computing, communication, and sensing—be documented with structured Model Cards analogous to those used for machine-learning models. The cards are meant to disclose what an entity is, what it is for, what quantum phenomena underpin it, what can go wrong, how its performance is measured, and what evidence supports those claims. The authors argue this transparency is needed for technology selection, systems integration, risk management, regulatory compliance, and roadmapping. They provide a detailed ten-section template and recommend that cards be encoded with FAIR metadata so they can be searched and mined automatically. The paper is explicitly a starting proposal: the final form would need community agreement and standardisation.","feed_headline":"Standardized model cards proposed for quantum technologies","feed_subtitle":"A ten-section form would spell out specs, errors, benchmarks, and assurance evidence for quantum computers, sensors, and QKD.","key_machinery":"The central object is the quantum Model Card template, a ten-section structured document: Entity Details; Intended Use; Factors; Quantum Technology Specifications; Errors; Performance Metrics; Ethical Considerations; Evaluation Criteria; Assurability; and Supplementary Materials. Its mechanism is forced disclosure: each section obliges a developer to state not only what a quantum entity is claimed to do, but also its limitations, residual errors, evaluation conditions, and supporting evidence. The framework's extensibility is carried by a recommendation that cards be created with metadata conforming to FAIR principles, with 18-word statements for purpose and use cases, and with supporting documents such as FMEA analyses and assurance cases.","core_discovery":"The paper's central claim is that the documentation framework known as Model Cards—originally developed for reporting machine-learning models—can be extended into a general, extensible reporting standard for quantum technologies. It defines the quantum 'entity' broadly enough to cover computers, communication systems, sensors, and sub-components, and it insists the cards must capture quantum-specific information that classical datasheets do not: coherence times, entanglement structure, error sources with non-classical failure modes, error budgets, benchmarks such as quantum volume and QBER, interface and control requirements, and assurance evidence such as certifications, audit reports, and security proofs. On the paper's account, such documentation would let end users evaluate a device against its stated use cases, let systems engineers judge fitness for integration, and let policy makers track the technology ecosystem. The claim is not that any current device meets a quality bar, but that a standardised reporting structure is a necessary step toward transparency and assurability.","pith_inferences":["One extension the paper leaves implicit is a direct test: fill in the proposed template for an existing quantum device and ask independent evaluators whether the card's claims are reproducible from the supplied measurement details.","If adoption is the goal, the template may work better as a minimal core card plus optional annexes; a too-long mandatory form could push vendors toward vague entries, which would defeat the transparency purpose.","The same card structure could in principle be extended to hybrid classical-quantum systems, since the Interface and Evaluation sections already expect classical components to be specified.","The paper's honesty premise could be tested empirically: if cards become common, check whether reported residual errors and limitations correlate with third-party audit results; if not, incentives beyond voluntary disclosure are needed."],"forward_implications":["If quantum Model Cards became standard, a systems engineer could compare two candidate devices on the same reported metrics, including residual errors and fundamental limits, before integrating either.","End users could track a deployed entity's performance against the card's stated use cases over time, surfacing drift or degradation.","Regulators and policy makers would gain a common format for assessing risk and compliance across different quantum technology types.","Machine-readable FAIR-compliant cards could feed automated roadmapping and trade-space exploration, linking device-level capability to ecosystem-level targets.","The reporting structure could also help the community build standardised taxonomies and layer models for quantum systems, since card fields would expose where current categories break down."],"supporting_citations":[{"why":"Provides the volumetric benchmarking framework that the Performance Metrics section can draw on for quantum computers.","marker":"[1]"},{"why":"Supplies the original Model Cards framework for model reporting that this proposal extends to quantum entities.","marker":"[2]"},{"why":"Documents the datasheets-for-datasets approach, supporting the argument for structured disclosure of context and limitations.","marker":"[3]"},{"why":"Records the semiconductor roadmapping effort invoked as evidence that coordinated documentation accelerates an industry.","marker":"[4]"},{"why":"Defines the FAIR data principles the paper recommends for encoding cards in machine-readable, self-describing metadata.","marker":"[5]"},{"why":"Supplies criteria for selecting performance metrics by high-priority requirements and risk priority numbers, used in the metrics section.","marker":"[8]"},{"why":"Provides the structured assurance case metamodel offered as a way to include assurance evidence in a card.","marker":"[9]"}],"fun_headline_variants":["Model cards for quantum tech: transparency blueprint","Quantum devices get AI-style model cards for reporting","Proposed standard: model cards for quantum technologies","Ten-section model cards to document quantum performance"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The framework's usefulness depends on stakeholders voluntarily providing accurate, complete, and honest information in model cards; the paper itself states that usefulness and accuracy depend on those engaging with it to act with transparency and include sufficient data on assurability, verification, and validation.","fun_headline_variants_meta":{"raw":{"variants":["Model cards for quantum tech: transparency blueprint","Quantum devices get AI-style model cards for reporting","Proposed standard: model cards for quantum technologies","Ten-section model cards to document quantum performance"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00034,"raw_usage":{"total_tokens":1827,"prompt_tokens":850,"completion_tokens":977,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":466,"completion_tokens_details":{"reasoning_tokens":920}},"tokens_in":466,"tokens_out":977,"duration_ms":7422,"temperature":1.0,"reasoning_tokens":920,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T13:21:54.295813+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take any existing quantum device—say, a quantum key distribution link or a small superconducting processor—fill in the proposed card from the developer's documentation, and have a second team independently reproduce the performance metrics by following the card's measurement and evaluation sections. If the card cannot be completed without recourse to undocumented assumptions, or if independent replication yields materially different numbers, the template fails as a transparency device. A simpler observational test: if model cards see widespread use and yet almost none report residual errors, limitations, or failed evaluations, the voluntary-honesty premise has not held.","supporting_citations":[{"cited_title":"This includes quantum gates, quantum opera- tions, and quantum algorithm’s tailored to the system’s capabilities","cited_arxiv_id":null,"evidence_quote":"Provides the volumetric benchmarking framework that the Performance Metrics section can draw on for quantum computers."},{"cited_title":"Circuit parame- ters for each component should be given","cited_arxiv_id":null,"evidence_quote":"Supplies the original Model Cards framework for model reporting that this proposal extends to quantum entities."},{"cited_title":"Data Handling: Formats and protocols for how the in- terfacing data processing systems handle the transmitted data","cited_arxiv_id":null,"evidence_quote":"Documents the datasheets-for-datasets approach, supporting the argument for structured disclosure of context and limitations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Records the semiconductor roadmapping effort invoked as evidence that coordinated documentation accelerates an industry."},{"cited_title":"Blume-Kohout and K","cited_arxiv_id":null,"evidence_quote":"Defines the FAIR data principles the paper recommends for encoding cards in machine-readable, self-describing metadata."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies criteria for selecting performance metrics by high-priority requirements and risk priority numbers, used in the metrics section."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the structured assurance case metamodel offered as a way to include assurance evidence in a card."}],"review_version":1}