Communication Dans Un Congrès Année : 2025

Robust Classification in Bayesian Neural Networks

Résumé

Bayesian neural networks (BNNs) are competitive multiclass classifiers with reasonably sound probabilistic characteristics. In this paper, we consider BNNs as classifier ensembles obtained through sampling, and define the inference problem in BNNs as 1. computing a median probability distribution from the ensemble of distributions, under a given distance, and 2. finding a Bayes-optimal prediction (BOP) under the evaluation metric given the median. With this (re)formulation, all the results related to computing medians of sets of probability distributions can be leveraged to strengthen the predictive performance of BNNs. We shall recall a generic formulation of the problem of computing medians and provide empirical evidence to illustrate the potential impact of the choice of distance regarding accuracy metrics and calibration errors.

Fichier principal
Vignette du fichier
IUKM 2025 Robust_Classification_in_Bayesian_Neural_Networks.pdf (501.84 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-05049687 , version 1 (28-04-2025)

Licence

Identifiants

Citer

Kim-Dung Tran, Xuan-Truong Hoang, Vu-Linh Nguyen, Sébastien Destercke, Van-Nam Huynh. Robust Classification in Bayesian Neural Networks. 11th International Symposium on Integrated Uncertainty in Knowledge Modelling and Decision Making (IUKM 2025), Mar 2025, Ho Chi MInh City, Vietnam. pp.29-41, ⟨10.1007/978-981-96-4606-7_3⟩. ⟨hal-05049687⟩
224 Consultations
209 Téléchargements

Altmetric

Partager

  • More