arXiv AI
Divide et Calibra: Multiclass Local Calibration via Vector Quantization
Read the original on arXiv AI →The Flow has not summarised this story yet — read it at arXiv AI.
Related stories
arXiv AI
Jun 2
CalArena: A Large-Scale Post-Hoc Calibration Benchmark
arXiv:2605. 30188v2 Announce Type: replace-cross Abstract: Reliable probability estimates are critical in many machine learning applications, yet modern classifiers are often poorly calibrated.
arXiv Machine Learning
Jul 7
Geometry-Aware Bayesian Quantification via Compositional Data Analysis
arXiv:2607. 04977v1 Announce Type: new Abstract: Accurately estimating the unknown target label distribution is the critical first step for adapting to label shift.
arXiv Machine Learning
Oct 9
V-ECE: Estimating General Expected Calibration Errors
The paper introduces V-ECE, a variational estimator for general Expected Calibration Errors (ECE) that avoids binning and works with any convex divergence, including Lp distances. By using prediction‑dependent proper scores and a residual‑fitted temperature scaling via gradient boosting, V‑ECE provides a lower‑bound estimate of true calibration error. Experiments on semi‑synthetic benchmarks show V‑ECE outperforms existing binary and multiclass estimators, with additional theoretical insights on bias and confidence estimation.
arXiv Computer Vision
Oct 23
Residual Calibration via Local Feature-Space Refinement
The paper introduces Local Adaptive Probability Estimation (LAPE), a post‑hoc residual calibration framework that refines global calibrator outputs using non‑parametric local evidence from neighboring samples in feature space. LAPE serves as a plug‑and‑play module that leverages feature‑space geometry and reuses calibration data, improving calibration quality for various baselines. The authors also extend the approach to multi‑view inputs with Cascade Feature Augmentation Fusion (CFAF) and demonstrate its effectiveness through extensive experiments.
arXiv AI
Jul 28
DecoupleMix: Decoupled Ratio Search and Convex Allocation for Scalable VLM Data Recipes
arXiv:2607. 24516v1 Announce Type: cross Abstract: While data curation for Vision Language Models (VLMs) is increasingly active, public practice for constructing pretraining mixtures remains largely heuristic: practitioners stack datasets that pass quality filters, set cross-domain ratios by intuition, and lack a principled, attributable criterion for admitting new data, while frontier recipes remain undisclosed.
arXiv Machine Learning
Oct 23
Is High Local Accuracy Enough? Geometric Consensus with Support-Adaptive Statistical Personalization for Federated Learning
The paper introduces FedGRC, a federated learning framework that enhances both global recognition and personalization for heterogeneous clients. It combines Consensus‑Guided Representation Learning, Multilayer Metric Synthesis, and Support‑Adaptive Shrinkage Personalization to address failures on rare or unseen classes. Experiments on CIFAR‑10, CIFAR‑100, and the FedRS dataset show that FedGRC outperforms 21 baseline methods, improving global accuracy and personalization floors by over 2% in several settings.