ZClassifier: Temperature Tuning and Manifold Approximation via KL Divergence on Logit Space
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 Kullback-Leibler (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 show that ZClassifier improves over softmax classifiers in robustness, calibration, and latent separation.
Code (1)
Tasks
Out of Distribution (OOD) DetectionMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Robust Bi-Tempered Logistic Loss Based on Bregman Divergences
We introduce a temperature into the exponential function and replace the softmax output layer of neural nets by a high temperature generalization. Similarly, the logarithm in the log loss we use for training is replaced …
Tethered Reasoning: Decoupling Entropy from Hallucination in Quantized LLMs via Manifold Steering
Quantized language models face a fundamental dilemma: low sampling temperatures yield repetitive, mode-collapsed outputs, while high temperatures (T > 2.0) cause trajectory divergence and semantic incoherence. We present…
A Graph-based approach to derive the geodesic distance on Statistical manifolds: Application to Multimedia Information Retrieval
In this paper, we leverage the properties of non-Euclidean Geometry to define the Geodesic distance (GD) on the space of statistical manifolds. The Geodesic distance is a real and intuitive similarity measure that is a g…
Information RetrievalRetrievalManifold Drift in Flow Preference Optimization: A Root Cause of Reward Hacking
Preference optimization is a standard alignment method for generative models, yet extending it to continuous-time dynamics remains non-trivial. In flow matching, reward-driven updates modify transport trajectories withou…
Two-temperature logistic regression based on the Tsallis divergence
We develop a variant of multiclass logistic regression that is significantly more robust to noise. The algorithm has one weight vector per class and the surrogate loss is a function of the linear activations (one per cla…
regressionVocal Bursts Valence Prediction