ZClassifier: Temperature Tuning and Manifold Approximation via KL Divergence on Logit Space
This work presents a novel method for uncertainty calibration and latent control in classification, though it is incremental as it builds on existing probabilistic frameworks.
The paper tackles classification by replacing deterministic logits with Gaussian-distributed logits to address temperature scaling and manifold approximation, resulting in improved robustness, calibration, and latent separation on CIFAR-10 and CIFAR-100 datasets.
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.