发表机构
Hanyang University; Hankuk University of Foreign Studies（汉阳大学; 韩国外国语大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对事后校准会改变Top-1预测的问题，提出CORD适配器，通过修复校准概率向量实现零预测改变率，同时降低ECE、NLL和Brier等指标，在多数据集及分布偏移下均有效。
AI 中文摘要
事后校准用于修正报告的置信度，但多分类校准器也会改变关联的Top-1预测。准确率仅捕捉这些改变对正确性的净效应，无法体现预测改变的频率；Top-1预测改变率（TPCR）则用于衡量该频率。我们提出Top-1决策保留的校准器输出修复方法（CORD），这是首个通过修复完整校准概率向量来强制实现精确预测保留的后适配适配器。仅利用原始输出和校准输出，CORD即可确定分配给原始Top-1类别的质量；校准后的条件分布将剩余质量分配给其他类别，生成的修复向量的argmax可恢复原始预测。在校准分割集上，CORD协调修复后的质量，尽可能保留校准输出在原始预测上的平均质量。该适配器既不改变已拟合的校准器及其直接输出，也无需拟合额外的监督映射，且无需用户或验证调整超参数。在CIFAR-10/100和ImageNet-1K数据集上，CORD通过构造实现零TPCR，且在所有数据集上相比对应直接输出均降低了平均预期校准误差（ECE）、负对数似然（NLL）和布里尔分数（Brier）；在分布偏移及不同校准集规模下，配对增益依然存在。因此，CORD在校准器拟合中移除了保留约束，将原始决策的精确恢复分配给后续输出修复。我们的代码可在该https URL获取。
英文摘要
Post-hoc calibration corrects reported confidence, yet a multiclass calibrator can also change the associated top-1 prediction. Accuracy captures only the net effect of these changes on correctness, not how often predictions change; the Top-1 Prediction Change Rate (TPCR) instead measures this frequency. We propose Calibrator-Output Repair for Top-1 Decision Preservation (CORD), the first post-fit adapter to impose exact prediction preservation by repairing the full calibrated probability vector. From the original and calibrated outputs alone, CORD determines the mass assigned to the original top-1. The calibrated conditional distribution allocates the remaining mass over the other classes, yielding a repaired vector whose own argmax recovers the original prediction. On the calibration split, CORD coordinates the repaired masses to retain the calibrated outputs' mean mass on original predictions whenever attainable. The adapter alters neither the fitted calibrator nor its direct output, fits no additional supervised map, and requires no user- or validation-tuned hyperparameter. Across CIFAR-10/100 and ImageNet-1K, CORD attains zero TPCR by construction and lowers mean ECE, NLL, and Brier relative to the corresponding direct outputs in every dataset; paired gains persist under distribution shift and across calibration-set sizes. CORD thus removes the preservation constraint from calibrator fitting and assigns exact recovery of the original decision to subsequent output repair. Our code is available at https://github.com/labhai/CORD.
