Saved in:
Bibliographic Details
Main Authors: Dengjin, Li, Yanming, Guo, Yuxiang, Xie, Zheng, Li, Jiangming, Chen, Xiaolong, Li, Mingrui, Lao
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.00089
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915470705164288
author Dengjin, Li
Yanming, Guo
Yuxiang, Xie
Zheng, Li
Jiangming, Chen
Xiaolong, Li
Mingrui, Lao
author_facet Dengjin, Li
Yanming, Guo
Yuxiang, Xie
Zheng, Li
Jiangming, Chen
Xiaolong, Li
Mingrui, Lao
contents Ensemble Adversarial Training (EAT) attempts to enhance the robustness of models against adversarial attacks by leveraging multiple models. However, current EAT strategies tend to train the sub-models independently, ignoring the cooperative benefits between sub-models. Through detailed inspections of the process of EAT, we find that that samples with classification disparities between sub-models are close to the decision boundary of ensemble, exerting greater influence on the robustness of ensemble. To this end, we propose a novel yet efficient Collaborative Ensemble Adversarial Training (CEAT), to highlight the cooperative learning among sub-models in the ensemble. To be specific, samples with larger predictive disparities between the sub-models will receive greater attention during the adversarial training of the other sub-models. CEAT leverages the probability disparities to adaptively assign weights to different samples, by incorporating a calibrating distance regularization. Extensive experiments on widely-adopted datasets show that our proposed method achieves the state-of-the-art performance over competitive EAT methods. It is noteworthy that CEAT is model-agnostic, which can be seamlessly adapted into various ensemble methods with flexible applicability.
format Preprint
id arxiv_https___arxiv_org_abs_2509_00089
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Learning from Peers: Collaborative Ensemble Adversarial Training
Dengjin, Li
Yanming, Guo
Yuxiang, Xie
Zheng, Li
Jiangming, Chen
Xiaolong, Li
Mingrui, Lao
Machine Learning
Ensemble Adversarial Training (EAT) attempts to enhance the robustness of models against adversarial attacks by leveraging multiple models. However, current EAT strategies tend to train the sub-models independently, ignoring the cooperative benefits between sub-models. Through detailed inspections of the process of EAT, we find that that samples with classification disparities between sub-models are close to the decision boundary of ensemble, exerting greater influence on the robustness of ensemble. To this end, we propose a novel yet efficient Collaborative Ensemble Adversarial Training (CEAT), to highlight the cooperative learning among sub-models in the ensemble. To be specific, samples with larger predictive disparities between the sub-models will receive greater attention during the adversarial training of the other sub-models. CEAT leverages the probability disparities to adaptively assign weights to different samples, by incorporating a calibrating distance regularization. Extensive experiments on widely-adopted datasets show that our proposed method achieves the state-of-the-art performance over competitive EAT methods. It is noteworthy that CEAT is model-agnostic, which can be seamlessly adapted into various ensemble methods with flexible applicability.
title Learning from Peers: Collaborative Ensemble Adversarial Training
topic Machine Learning
url https://arxiv.org/abs/2509.00089