AI 中文总结
该研究提出QuantumPhaseNet，将Transformer表示扩展为规范协变几何与量子谱模型，经内置合成设置验证，其在语义层级、话语对齐等任务上优于基线，但未展现量子优势。
AI 中文摘要
我们提出了QuantumPhaseNet，这是Transformer表示的一种规范协变几何与量子谱扩展模型。上下文相关的语义状态被建模为复振幅；协变相位率诱导出用作概念尺度代理的语义波长；低频图模式则定义了文档级的话语方向。理论部分确立了局域规范不变性、量子块的幺正性、WavePhase注意力的有界性与条件稳定性，以及可校准的幻觉风险公式。我们还在第14.1节中为经典量子启发流程实现了完全离线的Validation Studio，并在其内置合成设置（n=240，观测噪声0.22，电路噪声0.08，5个随机种子）上针对第16.1节中的5个研究问题（RQ）进行了评估。RQ1得出波长层级的斯皮尔曼相关系数为0.852，而基线为0.707，方向准确率为87.3%，AUC为0.953；RQ2的话语对齐度为0.933，而基线为0.589，漂移前段落数为41.2，而基线为16.2；RQ3的AUROC为0.881，而余弦方法为0.765、相位打乱方法为0.536；RQ4的错误检测AUROC为0.854，而熵方法为0.634，Brier值为0.150，ECE为0.098；RQ5未展现量子优势：目标概率与端到端成本效率分别为25.5%与0.107，而切比雪夫经典近似方法分别为70.7%与0.707。这些结果为经典量子启发组件提供了初步的合成证据，但未获得外部有效性或无条件量子加速。
英文摘要
We present QuantumPhaseNet, a gauge-covariant geometric and quantum-spectral extension of Transformer representations. Context-dependent semantic states are modeled as complex amplitudes; a covariant phase rate induces a semantic wavelength used as a proxy for conceptual scale; and low-frequency graph modes define a document-level discourse direction. The theoretical part establishes local gauge invariance, unitarity of the quantum block, boundedness and conditional stability of WavePhase Attention, and a calibratable hallucination-risk formulation. We also implemented a fully offline Validation Studio for the classical quantum-inspired pipeline in Section 14.1 and evaluated the five research questions in Section 16.1 on its built-in synthetic setting (n=240, observation noise 0.22, circuit noise 0.08, five seeds). RQ1 yielded a wavelength-hierarchy Spearman correlation of 0.852 versus 0.707 for the baseline, 87.3% direction accuracy, and AUC 0.953. RQ2 achieved discourse alignment 0.933 versus 0.589 and 41.2 versus 16.2 paragraphs before drift. RQ3 achieved AUROC 0.881 versus cosine 0.765 and phase-shuffle 0.536. RQ4 achieved error-detection AUROC 0.854 versus entropy 0.634, with Brier 0.150 and ECE 0.098. RQ5 did not show quantum advantage: target probability and end-to-end cost efficiency were 25.5% and 0.107, compared with 70.7% and 0.707 for the Chebyshev classical approximation. These results provide initial synthetic evidence for the classical quantum-inspired components, but not external validity or unconditional quantum speedup.
Comments[PAGES] pages, 8 figures, 4 tables. Extends arXiv:2602.14419 (WavePhaseNet). Includes prototype validation with an offline Validation Studio; RQ5 reports a negative result for quantum advantage
