AI-Shielder: Exploiting Backdoors to Defend Against Adversarial Attacks
- 1 January 2026
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Dependable and Secure Computing
- Vol. 23 (1) , 1244-1259
Abstract
No abstract availableKeywords
Funding Information
- Natural Science Foundation of Shandong Province (ZR2025QC1556)
- National Natural Science Foundation of China (62572285)
- National Natural Science Foundation of China (62032025)
- National Natural Science Foundation of China (62172258)
- National Natural Science Foundation of China (92270204)
- National Natural Science Foundation of China (U24A20236)
- National Natural Science Foundation of China (U21A20467)
- CAS Project for Young Scientists in Basic Research (YSBR-118)
This publication has 38 references indexed in Scilit:
- Collaborative Defense-GAN for protecting adversarial attacks on classification systemExpert Systems with Applications, 2022
- SemanticAdv: Generating Adversarial Examples via Attribute-Conditioned Image Editing2020
- Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks2020
- Square Attack: A Query-Efficient Black-Box Adversarial Attack via Random Search2020
- GeoDA: A Geometric Framework for Black-Box Adversarial Attacks2020
- Evading Defenses to Transferable Adversarial Examples by Translation-Invariant Attacks2019
- BadNets: Evaluating Backdooring Attacks on Deep Neural NetworksIEEE Access, 2019
- Deep Learning Face Representation from Predicting 10,000 Classes2014
- Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognitionNeural Networks, 2012
- Face recognition in unconstrained videos with matched background similarity2011