发表机构
Faculty of Data and Decision Sciences; Technion（数据与决策科学学院; 以色列理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出图平滑作为训练时校准的一般原则，通过惩罚相邻样本预测分布的Jensen-Shannon散度来改善神经网络校准，并在多个基准上验证了其与事后校准互补的有效性。
AI 中文摘要
现代神经网络常常校准不佳，倾向于过度自信。现有的训练时校准方法主要修改任务损失或校准惩罚，未能充分利用学习表示中的邻域结构。我们引入图平滑作为训练时校准的一般原则，它鼓励表示空间中相邻样本的预测分布相似。我们分析了图平滑的效果，推导出将相邻样本间的预测差异与局部置信度变化以及逐点校准误差的传播联系起来的界限，并刻画了平滑能够或不能改善校准的条件。基于这一分析，我们提出了\modelNoSpace，一种基于图的训练时正则化器，它惩罚相邻样本预测分布之间的Jensen-Shannon散度。我们进行了彻底的实证分析，表明在标准校准基准上，\model提高了预测质量，并且这种改进与事后校准互补：在温度缩放后，\model在八个图像和表格设置中的七个中取得了所有评估的训练时方法中最低的NLL。这些发现证明了图平滑对学习表示在神经网络校准中的价值。
英文摘要
Modern neural networks are often miscalibrated, with a tendency to overconfidence. Existing train-time calibration methods largely modify task losses or calibration penalties, leaving neighborhood structure in learned representations underexploited. We introduce graph smoothing as a general principle for train-time calibration, which encourages similar predictive distributions across neighboring samples in representation space. We analyze the effects of graph smoothing, deriving bounds that connect predictive divergence between neighboring samples to local confidence variation and to the propagation of pointwise calibration error, and characterize the conditions under which smoothing can or cannot improve calibration. In light of this analysis, we propose \modelNoSpace, a graph-based train-time regularizer that penalizes the Jensen--Shannon divergence between predictive distributions of neighboring samples. We present a thorough empirical analysis, showing that across standard calibration benchmarks, \model improves predictive quality, and the improvement is complementary to post-hoc calibration: after temperature scaling, \model attains the lowest NLL of all evaluated train-time methods in seven of the eight image and tabular settings. These findings demonstrate the value of graph smoothing over learned representations for neural network calibration.
