SOURCE-LINKED INTELLIGENCE

DEFUSE: Generalizable Backdoor Defense for Self-Supervised Encoders with Generative Priors

arXiv · AI, language, vision and robotics · article · Aug 26, 2026 · UTC

Self-supervised learning (SSL) encoders are vulnerable to backdoor attacks, posing threats to both visual SSL encoders and vision-language encoders. Existing defenses are typically designed for only one of these paradigms and rely on restrictive assumptions such as access to uninfected in-distribution data or precomputed pseudo-labels, which are difficult to satisfy in practice. To address these limitations, we propose DEFUSE, a generalizable backdoor detection framework for SSL encoders. Inspired by Bayesian posterior inference, we reformulate backdoor detection as a representation-conditione

Read original source ↗ Open in workspace

recordType

paper

region

Global