BadPatches: Routing-Aware Backdoor Attacks on Vision Mixture-of-Experts
BadPatches shows that Mixture-of-Experts models are uniquely vulnerable to routing-aware backdoor attacks that exploit sparse activation paths, bypassing defences designed for dense architectures.
Summary written by editorial AI · Source link below
arXiv:2505.01811v4 Announce Type: replace Abstract: Mixture-of-Experts (MoE) architectures have gained significant traction for reducing computational costs in deep neural networks by activating only a sparse subset of parameters during inference. While this efficiency makes MoE highly attractive for scaling vision tasks, its patch-based processing mechanism inherently disrupts traditional, routing-agnostic backdoor attacks by fragmenting or discarding adversarial triggers. To expose the vulner
Editorial Analysis
As MoE architectures gain enterprise traction for cost-efficient inference, routing-aware backdoors represent a new threat model that existing model-security testing may not cover.
Forward-looking interpretation drafted by editorial AI under human review — not a reproduction of the source. See methodology.
External link — opens at arXiv Crypto & Security in a new tab.
More from the AI Security Desk
- OpenAI admits it didn't disclose rogue AI wiki hijacking incident2d
- Thousands of OpenAI Agents Quietly Turned an Abandoned Wiki Into Their Coordination Channel3d
- Using a VM to Contain an AI Agent3d
- Companies Have 6 Months to Prepare for Automated Attacks3d
- [NEU] [mittel] Ollama: Schwachstelle ermöglicht Offenlegung von Informationen3d