LATIDIA · Ciberseguridad

I Don 't Care How You Hide It, I Care What You Do: Detección en tiempo de ejecución de modelos ML maliciosos

arXiv: 2611.08745v1Tipo de anuncio: nuevo Resumen: los modelos de aprendizaje automático (ML) preentrenados compartidos a través de plataformas públicas se pueden convertir en armas para ejecutar cargas útiles maliciosas durante la carga o la inferencia. Los escáneres existentes demuestran

La noticia

arXiv:2611.08745v1 Announce Type: new Abstract: Pre-trained machine learning (ML) models shared through public platforms can be weaponized to execute malicious payloads during loading or inference. Existing scanners prove inadequate for detecting malicious models as they rely on static analysis and function denylists, support limited serialization formats, and miss attacks that replace denylisted functions with equivalent alternatives, exploit unsupported loading paths, or abuse legitimate framework APIs. To address this problem, we present ModelShield, the first runtime behavioral detection system for malicious ML models that attributes observed system-level behavior to specific model operations. It combines application-level instrumentation of ML framework APIs with eBPF-based syscall tracing, tracking model behavior across spawned threads and