Minimax Bayesian Neural Networks

Robustness is an important issue in deep learning, and Bayesian neural networks (BNNs) provide means of robustness analysis, while the minimax method is a conservative choice in the classical Bayesian field. Recently, researchers have applied the closed-loop idea to neural networks via the minimax m...

Full description

Saved in:
Bibliographic Details
Main Authors: Junping Hong, Ercan Engin Kuruoglu
Format: Article
Language:English
Published: MDPI AG 2025-03-01
Series:Entropy
Subjects:
Online Access:https://www.mdpi.com/1099-4300/27/4/340
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850143795873579008
author Junping Hong
Ercan Engin Kuruoglu
author_facet Junping Hong
Ercan Engin Kuruoglu
author_sort Junping Hong
collection DOAJ
description Robustness is an important issue in deep learning, and Bayesian neural networks (BNNs) provide means of robustness analysis, while the minimax method is a conservative choice in the classical Bayesian field. Recently, researchers have applied the closed-loop idea to neural networks via the minimax method and proposed the closed-loop neural networks. In this paper, we study more conservative BNNs with the minimax method, which formulates a two-player game between a deterministic neural network and a sampling stochastic neural network. From this perspective, we reveal the connection between the closed-loop neural and the BNNs. We test the models on some simple data sets and study their robustness under noise perturbation, etc.
format Article
id doaj-art-bdd3c8ce9bf64dadb9dd40d4caf9466c
institution OA Journals
issn 1099-4300
language English
publishDate 2025-03-01
publisher MDPI AG
record_format Article
series Entropy
spelling doaj-art-bdd3c8ce9bf64dadb9dd40d4caf9466c2025-08-20T02:28:34ZengMDPI AGEntropy1099-43002025-03-0127434010.3390/e27040340Minimax Bayesian Neural NetworksJunping Hong0Ercan Engin Kuruoglu1Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, ChinaTsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, ChinaRobustness is an important issue in deep learning, and Bayesian neural networks (BNNs) provide means of robustness analysis, while the minimax method is a conservative choice in the classical Bayesian field. Recently, researchers have applied the closed-loop idea to neural networks via the minimax method and proposed the closed-loop neural networks. In this paper, we study more conservative BNNs with the minimax method, which formulates a two-player game between a deterministic neural network and a sampling stochastic neural network. From this perspective, we reveal the connection between the closed-loop neural and the BNNs. We test the models on some simple data sets and study their robustness under noise perturbation, etc.https://www.mdpi.com/1099-4300/27/4/340Bayesian neural networksrobustnessnoise perturbationminimax gameclosed-loop neural networksmaximal coding rate distortion
spellingShingle Junping Hong
Ercan Engin Kuruoglu
Minimax Bayesian Neural Networks
Entropy
Bayesian neural networks
robustness
noise perturbation
minimax game
closed-loop neural networks
maximal coding rate distortion
title Minimax Bayesian Neural Networks
title_full Minimax Bayesian Neural Networks
title_fullStr Minimax Bayesian Neural Networks
title_full_unstemmed Minimax Bayesian Neural Networks
title_short Minimax Bayesian Neural Networks
title_sort minimax bayesian neural networks
topic Bayesian neural networks
robustness
noise perturbation
minimax game
closed-loop neural networks
maximal coding rate distortion
url https://www.mdpi.com/1099-4300/27/4/340
work_keys_str_mv AT junpinghong minimaxbayesianneuralnetworks
AT ercanenginkuruoglu minimaxbayesianneuralnetworks