用概率化对数空间统一校准与潜在空间控制，提升分类鲁棒性。
ZClassifier: Temperature Tuning and Manifold Approximation via KL Divergence on Logit Space
- 将对数空间改为对角高斯分布，实现不确定性建模
- 在CIFAR-10/100上提升校准精度与类别分离能力
- 适合需要可信置信度和可控表示的分类任务
我们提出一种新型分类框架ZClassifier，将传统确定性对数空间替换为对角高斯分布的对数空间。通过最小化预测高斯分布与单位各向同性高斯之间的KL散度，该方法同时实现温度缩放与流形近似。这一统一的概率框架提升了置信度解释性与几何一致性。在CIFAR-10和CIFAR-100上的实验表明，ZClassifier在鲁棒性、校准性和潜在空间分离方面均优于Softmax分类器，且在小规模与大规模分类设置中均有稳定提升。
原文摘要 · Abstract (English)
We introduce a novel classification framework, ZClassifier, that replaces conventional deterministic logits with diagonal Gaussian-distributed logits. Our method simultaneously addresses temperature scaling and manifold approximation by minimizing the KL divergence between the predicted Gaussian distributions and a unit isotropic Gaussian. This unifies uncertainty calibration and latent control in a principled probabilistic manner, enabling a natural interpretation of class confidence and geometric consistency. Experiments on CIFAR-10 and CIFAR-100 demonstrate that ZClassifier improves over softmax classifiers in robustness, calibration, and latent separation, with consistent benefits across small-scale and large-scale classification settings.
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。