AI 中文总结
研究针对现代喷注标记器训练成本高且难共享问题，提出JetCoRD跨实验蒸馏方法，通过单个样本可靠性信号$r_i$创新设计，用仅82k可训练参数实现学生模型，在准确率上与教师模型相当且在特定工作点有提升，适用于压缩校准不完善教师模型场景。
AI 中文摘要
现代基于图和变压器网络的喷注标记器性能达到了先进水平，但训练成本高昂且难以在不同实验间共享。知识蒸馏原则上可实现模型压缩，但其基于教师模型是完美标记器的假设。在高纯度工作点该假设不成立，此时教师模型本身存在约10%到30%的喷注预测错误率。我们引入了JetCoRD，这是高能物理领域首个跨实验蒸馏方法：一个具有82k参数的统一学生模型，同时从ATLAS GN2（5M参数，3类）和CMS ParT（2M参数，10类）进行蒸馏。核心创新是单个样本可靠性信号$r_i$，它同时加权蒸馏损失、控制基于原型的教师修复，并锚定混合教师和学生逻辑的推理时门控$g_i$。仅82k可训练参数（占7M教师参数总和的1.2%），学生模型在总体准确率上与两个教师模型相当，在物理可操作工作点上超过它们：在$\varepsilon=0.77$时，$b$对$c$提高4.3%；在$\varepsilon=0.30$时，$c$对$u$提高1.5%；在$\varepsilon=0.5$时，$T\to bqq$提高1.6%；在$\varepsilon=0.5$时，$H\to bb$提高1.4%。可靠性耦合设计在高能物理蒸馏中是新颖的，适用于任何必须压缩校准不完善的教师模型的情况。
英文摘要
Modern jet taggers based on graph and transformer networks deliver state-of-the-art performance but are expensive to train and difficult to share across experiments. Knowledge distillation can in principle achieve model compression, but it rests on the assumption that the teacher model acts as a perfect tagger. This assumption fails to hold at high-purity working points, where the teacher itself exhibits a jet prediction error rate of approximately $10\%$ to $30\%$. We introduce JetCoRD, the first cross-experiment distillation in HEP: an 82 k-parameter unified student distilling jointly from ATLAS GN2 (5 M params, 3 classes) and CMS ParT (2 M params, 10 classes). The central innovation is a single per-sample reliability signal $r_i$ that simultaneously weights the distillation loss, controls prototype-based teacher repair, and anchors the inference-time gate $g_i$ that mixes teacher and student logits. With only 82 k trainable parameters (1.2% of the combined 7 M teacher parameters), the student matches both teachers in overall accuracy and exceeds them at physics-actionable working points: $+4.3\%$ on $b$-vs-$c$ at $\varepsilon=0.77$, $+1.5\%$ on $c$-vs-$u$ at $\varepsilon=0.30$, $+1.6\%$ on $T\!\to\!bqq$ at $\varepsilon=0.5$ and $+1.4\%$ on $H\!\to\!bb$ at $\varepsilon=0.5$. The reliability-coupled design is novel in HEP distillation and applicable wherever an imperfectly calibrated teacher must be compressed.
