提升多分类概率校准效果，让模型预测更可信。
Improving Multi-Class Calibration through Normalization-Aware Isotonic Techniques
- 提出两种新方法，自动考虑概率和为1的约束。
- 在多个数据集上，负对数似然和校准误差均显著降低。
- 适合需要高可靠性概率输出的场景，如医疗诊断。
准确可靠的概率预测对多分类监督学习至关重要，良好校准的模型可支持理性决策。尽管等倾回归在二分类中表现优异，但其通过一对多方式扩展到多分类时，相比参数化方法表现欠佳，限制了实际应用。本文提出新型等倾归一化感知校准技术，基于从业者自然直观的假设。与以往方法不同，我们的方法通过将归一化直接融入优化过程（NA-FIR）或建模为累积双变量等倾回归（SCIR），天然考虑概率和为1的约束。在多种文本与图像分类数据集、不同模型架构上的实证评估表明，该方法在负对数似然（NLL）和期望校准误差（ECE）上均有稳定提升。
原文摘要 · Abstract (English)
Accurate and reliable probability predictions are essential for multi-class supervised learning tasks, where well-calibrated models enable rational decision-making. While isotonic regression has proven effective for binary calibration, its extension to multi-class problems via one-vs-rest calibration produced suboptimal results when compared to parametric methods, limiting its practical adoption. In this work, we propose novel isotonic normalization-aware techniques for multiclass calibration, grounded in natural and intuitive assumptions expected by practitioners. Unlike prior approaches, our methods inherently account for probability normalization by either incorporating normalization directly into the optimization process (NA-FIR) or modeling the problem as a cumulative bivariate isotonic regression (SCIR). Empirical evaluation on a variety of text and image classification datasets across different model architectures reveals that our approach consistently improves negative log-likelihood (NLL) and expected calibration error (ECE) metrics.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。