We propose a post-hoc explanation layer for binary Quantum Logic Decision Tree classifiers constructed with QLDT+. The layer does not modify the fitted tree, the decision threshold, or the predicted label. It translates the class-labelled QLDT+ paths into IF–THEN rules and attaches an explanation-based confidence score to each prediction. This score is the equal-weight average of three quantities: rule support, the activation mass of the contributing rules of the predicted class; path reliability, a class-specific reliability derived from the training evidence ratio; and the normalized distance of the QLDT+ score from the decision threshold. In a stratified 5-fold cross-validation on the Banknote and Diabetes datasets, which was repeated with ten random seeds, correct predictions received a higher mean score than incorrect ones in every run (mean difference 0.18 and 0.14), and near-threshold decisions received lower scores. In selective prediction, the error decreased as coverage decreased; on Diabetes, however, the decisions retained at 40% coverage or less were all predicted as class 0. The layer is therefore best read as a decomposable explanation of a fixed QLDT+ decision rather than as a calibrated probability or an optimized selective-prediction score.
We propose a post-hoc explanation layer for binary Quantum Logic Decision Tree classifiers constructed with QLDT+. The layer does not modify the fitted tree, the decision threshold, or the predicted label. It translates the class-labelled QLDT+ paths into IF–THEN rules and attaches an explanation-based confidence score to each prediction. This score is the equal-weight average of three quantities: rule support, the activation mass of the contributing rules of the predicted class; path reliability, a class-specific reliability derived from the training evidence ratio; and the normalized distance of the QLDT+ score from the decision threshold. In a stratified 5-fold cross-validation on the Banknote and Diabetes datasets, which was repeated with ten random seeds, correct predictions received a higher mean score than incorrect ones in every run (mean difference 0.18 and 0.14), and near-threshold decisions received lower scores. In selective prediction, the error decreased as coverage decreased; on Diabetes, however, the decisions retained at 40% coverage or less were all predicted as class 0. The layer is therefore best read as a decomposable explanation of a fixed QLDT+ decision rather than as a calibrated probability or an optimized selective-prediction score.
The quantum-logic inspired decision tree (QLDT) is based on quantum logic concepts and input values from the unit interval whereas the traditional decision tree is based on Boolean values. The logic behind the QLDT obeys the rules of a Boolean algebra. The QLDT is appropriate for classification problems where for a class decision several input values interact gradually with each other. The QLDT construction for a classification problem with n input attributes requires the computation of 2n minterms. The QLDT+ method, however, uses a heuristic for obtaining a QLDT with much smaller computational complexity. As result, the QLDT+ method can be applied to classification problems with a higher number of input attributes.
Quantum error correction can be enhanced by post-selecting out runs that are likely to produce a logical failure, but the most accurate measures for that require costly decoder-level information. We introduce a practical, decoder-agnostic post-selection method that learns directly from syndrome data. The method trains a supervised classifier to distinguish between syndromes from low- and high-noise regimes, and then uses the classifier's output as an abort score for new runs, without requiring logical-error labels, correction operators, or code-specific likelihood calculations. We validate the approach in three complementary settings: circuit-level simulations of the Gross bivariate-bicycle code, code-capacity simulations of the surface code, and experimental logical magic-state distillation data from the QuEra neutral-atom processor. In the Gross and surface codes, learned syndrome post-selection reduces the conditional logical error rate at a fixed acceptance rate, with performance comparable to syndrome-weight filtering. For the surface code, the learned classifier reveals a post-selection transition distinct from the conventional decoding threshold. In the experimental data, the machine-learning score outperforms syndrome-weight post-selection and, when combined with logical-gap filtering, improves the output fidelity beyond using the logical gap alone. These results show that syndrome-only learning provides a scalable and hardware-compatible route to improving the reliability of quantum error correction.
Abstract: This study investigates an explainable quantum-logical multi-agent system (QL-MAS) for autonomous scientific discovery in high-dimensional research landscapes. The central methodological problem is that modern discovery settings combine large numbers of variables, heterogeneous evidence types, and partially conflicting reasoning signals, yet conventional single-model pipelines often compress this complexity into opaque rankings. The proposed framework addresses this limitation by treating candidate scientific hypotheses as quantum-logical superpositions of evidence generated by three complementary agents: an evidence agent based on mutual information, a structure agent based on principal-component loadings, and a rule agent based on interpretable decision structures. The resulting architecture does not claim physical quantum computation; rather, it uses quantum-logical amplitude composition as a mathematically disciplined way to represent concurrent support, interference, and consensus across agents. Empirical evaluation is conducted on three real public datasets that approximate heterogeneous scientific discovery landscapes: the Breast Cancer Wisconsin (Diagnostic) dataset (569 observations, 30 variables), the Wine Recognition dataset (178 observations, 13 variables), and the Diabetes dataset used in the least-angle-regression literature (442 observations, 10 baseline variables). These corpora were selected because they are non-synthetic, scientifically grounded, and sufficiently high-dimensional to test feature prioritization, interpretability, and cross-domain transfer. Across the three datasets, the proposed system identifies compact sets of high-salience hypotheses while preserving strong downstream predictive utility. The top-five features selected by QL-MAS attain cross-validated scores of 0.965 on breast cancer classification, 0.955 on wine classification, and 0.465 R² on diabetes progression, remaining close to full-feature baselines while offering substantially more interpretable evidence paths. The results indicate that explainable multi-agent discovery is strongest when high predictive relevance and rule-based consistency converge, and weakest when latent structure is diffuse or poorly aligned with supervised objectives. The analysis therefore supports the conclusion that autonomous discovery systems in research environments should not be assessed solely by predictive accuracy, but by the transparency of their evidential decomposition and by the stability of agreement across reasoning agents. The paper contributes a theoretically coherent framework, a real-data evaluation protocol, and a discussion of limitations, governance, and future directions for autonomous scientific reasoning systems. Keywords: autonomous scientific discovery; explainable AI; multi-agent systems; quantum logic; high-dimensional data; interpretable machine learning; research landscapes
To utilize quantum error-correcting codes, a decoder must infer the logical sector from the measured syndrome. Beyond a hard logical decision, some decoders provide soft information estimating its reliability. For minimum-weight perfect matching (MWPM), a common confidence measure is the complementary, or logical, gap. Neural-network decoders trained with binary cross entropy naturally output a different score, the logit, whose relation to the matching gap remains unclear. Here, we test whether the logit of a graph neural network (GNN) decoder behaves as a learned proxy for the logical gap. Using a pretrained GNN for the rotated surface code under uniform circuit-level noise [PRR, 7(2):023181, 2025], we compare its logit with the MWPM gap shot by shot on identical syndromes. The native GNN decoder-logit pair yields a lower post-selected logical error rate than the native MWPM decoder-gap pair over the operationally relevant range of acceptance rates. Exchanging the scores between decoders isolates score discrimination from hard-decision accuracy and shows that both contribute to the advantage. At larger distances each score best ranks failures of its own decoder, whereas at small distance, where the bounded matching gap saturates, the learned confidence better discriminates MWPM failures than the gap itself. Finally, GNN confidence is substantially better described by the posterior logistic form implied by an ideal log-likelihood ratio, while the MWPM gap shows systematic departures. Thus, a neural-network decoder trained only on syndromes and logical labels learns a gap-like confidence proxy whose native scale more closely follows the expected posterior calibration form, supporting learned soft output when MWPM gap estimates are unavailable, costly, or insufficiently expressive.
We introduce Decision Tree Decoders (DTDs), which rely only on the sparsity of the binary check matrix, making them broadly applicable for decoding any quantum low-density parity-check (qLDPC) code and fault-tolerant quantum circuits. DTDs construct corrections incrementally by adding faults one-by-one, forming a path through a Decision Tree (DT). Each DTD algorithm is defined by its strategy for exploring the tree, with well-designed algorithms typically needing to explore only a small portion before finding a correction. We propose two explicit DTD algorithms that can be applied to any qLDPC code: (1) A provable decoder: Guaranteed to find a minimum-weight correction. While it can be slow in the worst case, numerical results show surprisingly fast median-case runtime, exploring only w DT nodes to find a correction for weight-w errors in notable qLDPC codes, such as bivariate bicycle and color codes. This decoder may be useful for ensemble decoding and determining provable code distances, and can be adapted to compute all minimum-weight logical operators of a code. (2) A heuristic decoder: Achieves higher accuracy and faster performance than BP-OSD on the gross code with circuit noise in realistic parameter regimes.