发表机构
Query Machines B.V.（查询机器有限公司）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究通过量子Transformer块中的互信息、纠缠熵等物理量实现内在可解释性，实验证明纠缠是机制核心，并在IBM量子硬件上验证了推理过程的可观测性。
AI 中文摘要
深度学习模型功能强大但难以解释。随着量子机器学习的成熟，该领域面临一个决定性选择：构建同样难以解释的量子模型，或者利用量子力学的数学结构使其具有内在可解释性。我们证明后者是可能的。通过跟踪量子互信息（MI）、纠缠熵和态保真度在量子Transformer块（QTB）各层中的变化——QTB是一种具有注意力和前馈的量子类比的完全相干变分电路——我们直接洞察模型如何处理信息：它关注哪些令牌，相关性何时形成，以及预测为何失败。在四个具有已知依赖结构的任务上，我们表明：（i）学习到的MI矩阵与真实任务结构对齐（查找任务上AUC=0.69）；（ii）禁用纠缠门将准确率从100%降至15%，同时MI趋近于0，证明纠缠是机制所在；（iii）训练过程中准确率和MI共同演化（查找任务上ρ=0.92）；（iv）在条件任务上，每样本MI以ROC AUC=0.84预测预测正确性。所有结果均在IBM量子硬件（ibm_kingston，Heron r2）上验证：电路的推理过程，从乘积态到结构化纠缠，在超导处理器上直接可观测。这些在小型合成任务上获得的概念验证结果表明，量子计算的物理特性可以提供无直接经典对应物的内在可解释性信号，激励研究这种优势在更大规模上是否持续。
英文摘要
Deep learning models are powerful but opaque. As quantum machine learning matures, the field faces a defining choice: build quantum models that are equally opaque, or exploit the mathematical structure of quantum mechanics to make them inherently interpretable. We show that the latter is possible. By tracking quantum mutual information~(MI), entanglement entropy, and state fidelity through the layers of a Quantum Transformer Block (\qtb{}), a fully-coherent variational circuit with quantum analogues of both attention and feedforward, we gain direct insight into how the model processes information: which tokens it attends to, when correlations form, and why predictions fail. On four tasks with known dependency structure we show that (i)~learned MI matrices align with ground-truth task structure (AUC$\,{=}\,0.69$ on lookup), (ii)~disabling entangling gates collapses accuracy from 100\% to 15\% while MI$\to 0$, proving entanglement is the mechanism, (iii)~accuracy and MI co-evolve during training ($ρ\,{=}\,0.92$ on lookup), and (iv)~per-sample MI predicts prediction correctness on the conditional task with ROC AUC$\,{=}\,0.84$. All results are validated on IBM Quantum hardware (ibm\_kingston, Heron~r2): the circuit's reasoning process, from product state through structured entanglement, is directly observable on a superconducting processor. These proof-of-concept results, obtained on small synthetic tasks, suggest that the physics of quantum computation can provide intrinsic interpretability signals with no direct classical counterpart, motivating study of whether this advantage persists at scale.
