AI 中文总结
提出参考尾信任框架，在冻结GNN内进行有界学习更新并认证预测，通过概率下限和预算控制，实现可认证的GNN适应，优于现有校正器。
AI 中文摘要
图神经网络（GNN）需要在部署中不放弃对已有预测的控制的前提下，利用改进的消息传递。我们引入了参考尾信任（RTT），这是一个框架，允许在冻结的GNN内部进行学习更新，并认证实际服务的预测。RTT将基于图的提议状态与受约束的内部优化器相结合：每次位移都因其通过现有模型剩余消息传递层的最坏情况终端交叉熵增加而收费。轨迹验证管和独立检查器强制执行逐节点概率下限，$p^{\mathrm{s}}_{ic} \ge e^{-H_{\mathrm{row}}} p^{\mathrm{r}}_{ic}$，以及调用级预算，$\sum_i w_i D_\infty(p^{\mathrm{r}}_i \\| p^{\mathrm{s}}_i) \le H^+$，统一作用于标签。通过认证的适应输出的调用无需单独的完整现有模型展开；失败的证书触发整个调用的回退。我们通过水填充推导出精确的概率下限边界，刻画了架构受限的效率，并建立了内部传播利用受限输出校正器无法获得的证据的条件。在报告的ogbn-arxiv审计中，RTT每次调用实现了$6.5\times 10^{-3}$纳特的平均增益，在保留节点群体的未检查部分上，单侧95%回归率上限为0.95%，95%负翻转上限为0.51%。其平均增益是交叉拟合后验边界估计的61%，并超过最强匹配的单遍校正器$+0.9\times 10^{-3}$纳特。报告的实验涵盖八个提议、六个图现有模型家族、结构和时间图转移以及分子预测，并附有额外的图像和表格评估。RTT使GNN适应成为有预算的、可认证的推理决策，而非无条件的模型替换。
英文摘要
Graph neural networks (GNNs) need to exploit improved message passing without surrendering control over predictions already trusted in deployment. We introduce Reference-Tail Trust (RTT), a framework that admits learned updates inside a frozen GNN and certifies the prediction actually served. RTT couples graph-based proposal states with a constrained internal optimizer: each displacement is charged for its worst-case terminal cross-entropy increase through the incumbent's remaining message-passing layers. A trajectory-validated tube and an independent checker enforce per-node probability floors, $p^{\mathrm{s}}_{ic} \ge e^{-H_{\mathrm{row}}} p^{\mathrm{r}}_{ic}$, and a call-level budget, $\sum_i w_i D_\infty(p^{\mathrm{r}}_i \| p^{\mathrm{s}}_i) \le H^+$, uniformly over labels. Calls whose adapted outputs pass certification require no separate full incumbent rollout; failed certificates trigger whole-call fallback. We derive the exact probability-floor frontier by water-filling, characterize architecture-constrained efficiency, and establish conditions under which internal propagation exploits evidence unavailable to restricted output correctors. In the reported ogbn-arxiv audit, RTT achieves $6.5\times 10^{-3}$ nats of mean gain per call, with a one-sided 95% regression-rate upper bound of 0.95% and a 95% negative-flip upper bound of 0.51% on the uninspected part of the reserved node population. Its mean gain is 61% of a cross-fitted posterior-based frontier estimate and exceeds the strongest matched one-pass corrector by $+0.9\times 10^{-3}$ nats. Reported experiments span eight proposals, six graph-incumbent families, structural and temporal graph shifts, and molecular prediction, with additional image and tabular evaluations. RTT makes GNN adaptation a budgeted, certifiable inference decision rather than an unconditional model replacement.
Comments37 pages, 10 figures
