Ensemble Defense

Ensemble defense methods aim to improve the robustness and security of machine learning models by combining predictions from multiple individual models. Current research focuses on enhancing the diversity and effectiveness of these ensembles, exploring techniques like median-based weighting, novel aggregation strategies (e.g., two-round voting), and methods to reduce the transferability of adversarial attacks between ensemble members. These advancements are crucial for mitigating vulnerabilities to adversarial examples and data poisoning, thereby increasing the reliability and trustworthiness of machine learning systems in various security-sensitive applications.

Papers

October 15, 2024

Adversarially Guided Stateful Defense Against Backdoor Attacks in Federated Deep Learning
Hassan Ali, Surya Nepal, Salil S. Kanhere, Sanjay Jha
Backdoor Attack Adversarial Perturbation Federated Deep Ensemble Defense State Adversarial

June 20, 2024

MEAT: Median-Ensemble Adversarial Training for Improving Robustness and Generalization
Zhaozhe Hu, Jia-Li Yin, Bin Chen, Luojun Lin, Bo-Hao Chen, Ximeng Liu
Native Robustness Strong Generalization Adversarial Training Ensemble Adversarial Ensemble Defense

January 20, 2024

CARE: Ensemble Adversarial Robustness Evaluation Against Adaptive Attackers for Security Applications
Hangsheng Zhang, Jiqiang Liu, Jinsong Dong
Adversarial Robustness Security Application Ensemble Adversarial Ensemble Defense

February 5, 2023

Run-Off Election: Improved Provable Defense against Data Poisoning Attacks
Keivan Rezaei, Kiarash Banihashem, Atoosa Chegini, Soheil Feizi
Data Poisoning Attack Data Poisoning Defense Method Provable Defense Ensemble Defense

November 22, 2022

Self-Ensemble Protection: Training Checkpoints Are Good Data Protectors
Sizhe Chen, Geng Yuan, Xinwen Cheng, Yifan Gong, Minghai Qin, Yanzhi Wang, Xiaolin Huang
DNN Training Model Checkpoint Model Gradient Ensemble Defense Intermediate Checkpoint

November 15, 2022

MORA: Improving Ensemble Robustness Evaluation with Model-Reweighing Attack
Yunrui Yu, Xitong Gao, Cheng-Zhong Xu
Adversarial Attack Model Attack Robust Ensemble Ensemble Defense

October 6, 2022

Dynamic Stochastic Ensemble with Adversarial Robust Lottery Ticket Subnetworks
Qi Peng, Wenlin Liu, Ruoxi Qin, Libin Hou, Bin Yan, Linyuan Wang
Adversarial Attack Ensemble Defense Dynamic Ensemble Robust Adversarial

September 28, 2022

On the Robustness of Random Forest Against Untargeted Data Poisoning: An Ensemble-Based Approach
Marco Anisetti, Claudio A. Ardagna, Alessandro Balestrucci, Nicola Bena, Ernesto Damiani, Chan Yeob Yeun
Machine Learning Native Robustness Random Forest Poisoning Attack Random Forest Model Ensemble Based Ensemble Defense

August 18, 2022

Resisting Adversarial Attacks in Deep Neural Networks using Diverse Decision Boundaries
Manaar Alam, Shubhajit Datta, Debdeep Mukhopadhyay, Arijit Mondal, Partha Pratim Chakrabarti
Deep Neural Network Adversarial Attack Adversarial Example Adversarial Perturbation Decision Boundary Deep Learning System Ensemble Defense

June 11, 2022

Defending Adversarial Examples by Negative Correlation Ensemble
Wenjian Luo, Hongwei Zhang, Linghao Kong, Zhijian Chen, Ke Tang
Adversarial Example Adversarial Robustness DNN Framework Ensemble Defense Negative Correlation

December 7, 2021

Saliency Diversified Deep Ensemble for Robustness to Adversaries
Alex Bogun, Dimche Kostadinov, Damian Borth
Native Robustness Saliency Map Deep Ensemble Transferable Attack Adversary Agent Robust Ensemble Ensemble Defense