{"id":"70d0f4c5-8aa6-429a-bb11-1c2976ed9a77","arxiv_id":"2505.02033","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":4,"one_line_summary":"A 15-qubit deep variational quantum classifier trained on brain tumor microarray data attains about 85 percent cross-validated accuracy, comparable to some classical models but below SVM and random forest.","lead":"This paper applies a quantum machine learning model, a deep variational quantum classifier, to classify four brain tumor types and healthy samples from DNA microarray data. The model reaches roughly 85 percent cross-validated accuracy, which is similar to several classical algorithms but lower than support vector machines and random forests.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"15-qubit amplitude encoding cannot represent 54,676 features, so the full-feature experiment is not executable as described and the central accuracy claim is unsupported.","rationale":"The reader's REJECT verdict is well-founded. The strongest claim is about classifying from 54,676 gene features, and this is directly contradicted by the stated 15-qubit amplitude-encoding dimension. This is an internal inconsistency, not a disagreement with consensus: Eq. (1) itself limits the input length to 2^n, and 2^15 is 32,768. Every reported full-feature accuracy number is therefore tied to an impossible circuit as described. Even a generous reading that amplitude encoding could be approximated with truncation or padding has no textual basis, and the PCA variant does not rescue the abstract's full-feature claim because it uses a different input pipeline. I also share the reader's secondary concern that the CuMiDa baseline accuracies may not come from the same cross-validation partitions, which further weakens the comparison; however, the encoding mismatch alone is decisive. No code or data are supplied, so the discrepancy cannot be resolved by inspection. Since the reader already rejected the paper and my stress-test identifies the same decisive flaw, I set verdict_should_be to UNCHANGED.","tokens_in":8126,"tokens_out":3729,"duration_ms":41478,"concrete_test":"Re-run the exact Section II pipeline in PennyLane (default.qubit, 15 qubits) and call qml.AmplitudeEmbedding(features=vector_of_length_54676, wires=range(15), normalize=True). If the framework raises a shape/power-of-two error, the full-feature experiment is not reproducible without an undocumented truncation or a different number of qubits. As a secondary check, request the exact 3-fold split and seed used for the Deep VQC and rerun the CuMiDa classifiers on the same partitions to verify whether the Table I comparison is protocol-matched.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central experimental claim rests on a dimension mismatch. Section II.B defines amplitude encoding via Eq. (1) for an input vector of length 2^n. With n=15, the accessible amplitude space has 2^15 = 32,768 components. Section II.A.2 says the normalized microarray data were 'directly fed' into the Deep VQC model 'while preserving the original 54,676 gene features,' and Section III.A reiterates that all 54,676 features were transferred to the 15-qubit model. No truncation, padding, feature selection, or alternative encoding is described anywhere in the paper. Therefore the full-feature run, which produces the 0.79 validation accuracy and underlies the abstract's '54,676 gene features' claim, cannot be instantiated on the stated circuit. The PCA-reduced variant (65 features) is representable, but it uses a different input pipeline and cannot validate the headline full-feature result. A secondary protocol concern is that the CuMiDa ML baseline scores in Table I may come from a different held-out partition; even if the comparison were fair, the quantum model trails SVM, RF, and KNN, so 'quantum-enhanced' is an overreach. The encoding mismatch alone is sufficient to reject the central claim as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proposes a hybrid quantum-classical classifier called Deep VQC, which uses amplitude encoding, two hardware-efficient ansatze, and 15 qubits to classify five classes (four brain tumor types plus healthy samples) from CuMiDa microarray data with 54,676 gene features. The authors report a validation accuracy of 0.79 for the full-feature run, 0.86 when PCA is applied (reducing to 65 dimensions), and an average 3-fold cross-validation accuracy of 0.85. They compare these results with published CuMiDa classical ML baselines and conclude that the quantum model achieves competitive or superior performance.","tokens_in":8394,"tokens_out":2956,"duration_ms":31320,"significance":"If the claims were sound, the paper would provide a useful data point on applying variational quantum classifiers to high-dimensional genomic data, with a clear description of a circuit architecture and use of a public benchmark. The paper honestly reports learning curves and confusion matrices, and it cites the public CuMiDa database, which supports reproducibility of the dataset. However, the central claim is undermined by a direct encoding capacity mismatch: the full-feature experiment as described cannot be implemented on the stated 15-qubit circuit. In addition, the classical comparison uses published scores from a different protocol rather than same-protocol runs, and no error bars or repeated-run statistics are provided. These issues make the headline accuracy claims unsupported and the 'quantum-enhanced' framing an overreach.","major_comments":[{"comment":"The full-feature experiment is not executable as described. Eq. (1) defines amplitude encoding for an input vector of length 2^n, so with n=15 the state space has only 2^15 = 32,768 amplitude slots. Section II.A.2 states that normalized data were 'directly fed' into the Deep VQC model 'while preserving the original 54,676 gene features,' and Section III.A reiterates that all 54,676 features were transferred to the 15-qubit model. No truncation, feature selection, padding, or alternative encoding is described for the full-feature run. Therefore the reported 0.79 validation accuracy for this configuration cannot be reproduced on the stated hardware, and the abstract's '54,676 gene features' claim is unsupported.","section":"Section II.B.1, Eq. (1); Section II.A.2; Section III.A"},{"comment":"The comparison with classical ML models is not made under a common protocol. Table I reports CuMiDa database accuracy scores for classical algorithms, while the Deep VQC result comes from the authors' own 3-fold cross-validation run with a different (and unspecified) split, preprocessing, and hyperparameter selection. Without identical train/test partitions, feature scaling, and evaluation procedure, the accuracy values are not directly comparable. Moreover, even taking Table I at face value, the quantum model's 0.85 is lower than SVM (0.95), RF (0.91), and KNN (0.87), which contradicts the abstract's claim of 'superior or comparable' performance relative to classical ML algorithms.","section":"Section III.C and Table I"},{"comment":"No statistical uncertainty is reported for any of the quantum results. With a dataset of only 130 samples, single 3-fold cross-validation runs can have high variance, and the precision/recall/F1 ranges (e.g., recall from 0.25 to 1 for the full-feature model) indicate that some classes are poorly recognized. The paper does not provide standard deviations, confidence intervals, or repeated-run results, so the reported differences between settings (e.g., 0.79 vs. 0.86 validation accuracy) and the comparison with classical baselines are not statistically grounded.","section":"Section III.A and III.B"}],"minor_comments":[{"comment":"The number of PCA components is not stated in the preprocessing section; it first appears as '65 dimensions' in the conclusion. The methods should specify how many components were retained when keeping 95% of the variance.","section":"Section II.A.2 / Conclusion"},{"comment":"The denominator in Eq. (3) is difficult to read: the summation notation appears as 'P5' rather than a proper sum over k. Please clarify the formula.","section":"Eq. (3)"},{"comment":"The paper describes two hardware-efficient ansatze with different rotation gates, but Fig. 2 only shows the first layer and does not clearly depict both HEA structures or the mapping of the 15 qubits to the readout of five classes. A complete circuit diagram for both ansatze would improve reproducibility.","section":"Section II.B.2 and Fig. 2"},{"comment":"The phrase 'quantum-enhanced' is not justified by any comparison showing a quantum advantage over classical methods; the reported results are mostly comparable or worse than the classical baselines listed in Table I. The language should be softened to reflect a feasibility study.","section":"Abstract / Conclusion"}],"recommendation":"reject","confidential_remarks":"The encoding-capacity mismatch is the primary reason for rejection. The manuscript's central experimental claim—classifying from all 54,676 features with a 15-qubit amplitude-encoded circuit—cannot be instantiated as written. Even if the authors added a feature-selection step, the result would be a different experiment, and the classical comparison would still need to be re-run under a matched protocol. The paper is not suitable for publication in its current form."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know this paper in one line: the headline result is built on a dimension mismatch, so the full-feature accuracy claim cannot be reproduced as written. The authors say they feed all 54,676 gene features into a 15-qubit amplitude encoding, but Eq. (1) defines amplitude encoding for 2^n amplitudes, and 2^15 = 32,768. They never describe truncation, padding, or an alternative encoding. That makes the 0.79 validation accuracy for the full-feature run uninstantiable.\n\nWhat is actually new: applying a variational quantum classifier with two hardware-efficient ansatze and 25 layers to a brain-tumor microarray dataset is a new application. The PCA-reduced variant (65 features) is representable and runs, and the paper reports precision/recall/F1 ranges and a confusion matrix. That part is a legitimate, if incremental, benchmark. The claim that the model is 'quantum-enhanced' is not supported by their own comparison table: the Deep VQC gets 0.85, while SVM gets 0.95, RF 0.91, and KNN 0.87. So the model does not beat classical baselines, and the comparison uses CuMiDa's published scores rather than same-protocol runs, which weakens even the 'comparable' phrasing.\n\nThe soft spots are proportionate: the encoding mismatch is load-bearing and fatal to the central claim. The lack of error bars on the 0.85 cross-validation average is a lesser issue, but minor in comparison. No code or data are provided, so there is no way to check the PCA experiment either.\n\nWho is this for? A reader looking for a worked example of a VQC applied to a small expression dataset might get some value from the PCA section, but the paper as a whole does not support its abstract's claim of quantum-enhanced classification. It deserves a serious referee because the dimension mismatch is a concrete, checkable issue and the application is not absurd; but it needs major revision or a corrected claim before it is publishable.\n\nRecommendation: send to peer review with a request to verify the encoding step, or desk reject on the grounds of the internal contradiction.","headline":"The full-feature result is uninstantiable: 15 qubits cannot amplitude-encode 54,676 features, and the classical comparison does not support 'quantum-enhanced'.","tokens_in":8885,"tokens_out":1949,"would_cite":false,"duration_ms":18996,"reading_group":"no","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A 15-qubit variational quantum classifier labels four brain tumor types and healthy tissue from gene expression with about 85% accuracy.","keywords":["quantum machine learning","variational quantum classifier","brain tumor classification","DNA microarray","gene expression","amplitude encoding","hardware efficient ansatz","NISQ"],"falsifier":"Write out the 15-qubit state vector: amplitude encoding accepts at most $2^{15}=32{,}768$ amplitudes. Trying to load a 54,676-dimensional vector therefore fails unless the input is truncated, padded, or otherwise reduced; checking how the reported full-feature run handled this settles whether the result is reproducible. Re-running the classical baselines on the same three-fold splits would settle whether the accuracy comparison is fair.","tokens_in":7946,"feed_emoji":"🧠","tokens_out":8769,"duration_ms":79705,"temperature":0.7,"pith_summary":"The paper aims to show that a variational quantum classifier, named Deep VQC, can sort DNA microarray samples into five classes—four brain tumor types and healthy tissue—with useful accuracy. This matters because gene-expression data are extremely high-dimensional, with 54,676 features per sample, and classical methods face growing computational costs in this regime. The authors report validation accuracy of 0.79 when all 54,676 features are fed directly into the model, 0.86 after principal component analysis reduces the input to 65 dimensions, and an average of 0.85 under three-fold cross-validation. They also argue the quantum model is competitive with classical machine-learning baselines on the same dataset.","feed_headline":"15-qubit quantum model labels brain-tumor gene profiles at 85%","feed_subtitle":"A variational quantum classifier matches several classical machine-learning methods on five-class microarray brain tumor data.","key_machinery":"The workhorse is the Deep VQC, built from amplitude encoding, two hardware-efficient ansatze, and a softmax readout. Amplitude encoding expresses a normalized classical vector $x$ of length $2^n$ as the amplitudes of an $n$-qubit state, so a 15-qubit circuit can in principle carry at most $2^{15}=32{,}768$ amplitude components. The first HEA applies Hadamard gates, single-qubit RX and RY rotations, and CNOT and Toffoli entangling gates; the second uses Hadamard, RY and RZ rotations with the same entanglers. Measurement in the Pauli-Z basis on five qubits yields class probabilities through the softmax in Eq. (3), and the cross-entropy cost in Eq. (4) is minimized by gradient descent. This combination is what lets the model learn correlations among the gene-expression features while keeping the circuit shallow enough for NISQ hardware.","core_discovery":"The paper's central claim is that its Deep VQC model, a variational quantum classifier with 15 qubits and 25 layers, can separate five classes—ependymoma, glioblastoma, medulloblastoma, pilocytic astrocytoma, and healthy samples—from DNA microarray gene-expression data. The authors report validation accuracy 0.79 when all 54,676 features are transferred into the quantum model, 0.86 after principal component analysis reduces the input to 65 dimensions, and 0.85 average accuracy under three-fold cross-validation. They compare these numbers with classical machine-learning baselines and find the quantum model matches decision trees and naive Bayes, beats a multilayer perceptron, and trails support vector machines and random forests. On the paper's own account, the result is evidence that quantum AI can give competitive or better classification on high-dimensional biological data in the NISQ era.","pith_inferences":["The encoding formula in the paper accepts at most $2^{15}=32{,}768$ amplitudes, so the reported run with 54,676 features implies some unstated truncation, padding, or feature-reduction step; identifying that step would make the full-feature result reproducible.","The classical comparison is meaningful only if the same cross-validation splits and preprocessing were used; the paper does not document them, so the ranking against decision trees and naive Bayes should be read with that caveat.","The same two-ansatz Deep VQC design could be tried on other high-dimensional biological data, such as RNA-seq or methylation arrays, to test whether the pattern generalizes beyond DNA microarrays."],"forward_implications":["If the reported accuracy holds, a variational classifier with only 15 qubits is enough for a five-class medical diagnosis task on gene-expression data, so quantum classifiers are within reach of current hardware.","The gain from PCA (validation accuracy 0.86 over 0.79 on full features) suggests aggressive but variance-preserving dimensionality reduction can help a quantum classifier on high-dimensional biological data.","Matching decision trees and naive Bayes while beating a multilayer perceptron implies quantum variational classifiers are not automatically worse than classical methods on this kind of task.","The 25-layer two-ansatz structure indicates that circuit depth, not just qubit count, can be used as a resource in variational classifiers for microarray data."],"supporting_citations":[{"why":"Reports a prior quantum approach for distinguishing two brain tumor types, forming the application baseline this work extends.","marker":"[15]"},{"why":"Supplies the brain-tumor microarray dataset used for training and the classical machine-learning accuracy numbers used as comparison baselines.","marker":"[16]"},{"why":"Provides the variational quantum algorithm framework that defines the hybrid training loop used by Deep VQC.","marker":"[17]"},{"why":"Gives the amplitude-encoding formalism the paper relies on to load gene expression data into quantum states.","marker":"[18]"},{"why":"Establishes the hardware-efficient ansatz and NISQ context used for the two HEA circuits.","marker":"[19]"},{"why":"Provides the layered HEA expression the second ansatz is built upon.","marker":"[20]"}],"fun_headline_variants":["Quantum AI classifies brain tumors from gene data at 85%","15-qubit quantum model rivals classical ML on tumor gene profiles","Variational quantum classifier hits 85% on five brain tumor types","Deep VQC: quantum approach competes with classical ML for tumor classification","Quantum classifier reads 54k genes to label brain tumors with 85% accuracy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole result rests on the assumption that all 54,676 gene expression values can be loaded directly into a 15-qubit amplitude-encoding circuit, which the paper's own encoding formula does not allow at that size.","fun_headline_variants_meta":{"raw":{"variants":["Quantum AI classifies brain tumors from gene data at 85%","15-qubit quantum model rivals classical ML on tumor gene profiles","Variational quantum classifier hits 85% on five brain tumor types","Deep VQC: quantum approach competes with classical ML for tumor classification","Quantum classifier reads 54k genes to label brain tumors with 85% accuracy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00022,"raw_usage":{"total_tokens":1473,"prompt_tokens":996,"completion_tokens":477,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":612,"completion_tokens_details":{"reasoning_tokens":383}},"tokens_in":612,"tokens_out":477,"duration_ms":5309,"temperature":1.0,"reasoning_tokens":383,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T04:03:08.530748+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Write out the 15-qubit state vector: amplitude encoding accepts at most $2^{15}=32{,}768$ amplitudes. Trying to load a 54,676-dimensional vector therefore fails unless the input is truncated, padded, or otherwise reduced; checking how the reported full-feature run handled this settles whether the result is reproducible. Re-running the classical baselines on the same three-fold splits would settle whether the accuracy comparison is fair.","supporting_citations":[{"cited_title":"The role of quantum- enhanced support vector machine using multiparametric mri parameters in differentiating medulloblastoma from ependymoma,","cited_arxiv_id":null,"evidence_quote":"Reports a prior quantum approach for distinguishing two brain tumor types, forming the application baseline this work extends."},{"cited_title":"Cumida: An extensively curated microarray database for benchmarking and testing of machine learning approaches in cancer research,","cited_arxiv_id":null,"evidence_quote":"Supplies the brain-tumor microarray dataset used for training and the classical machine-learning accuracy numbers used as comparison baselines."},{"cited_title":"Variational quantum algorithms,","cited_arxiv_id":null,"evidence_quote":"Provides the variational quantum algorithm framework that defines the hybrid training loop used by Deep VQC."},{"cited_title":"Quantum chemistry in the age of quantum computing,","cited_arxiv_id":null,"evidence_quote":"Provides the layered HEA expression the second ansatz is built upon."}],"review_version":1}