LATIDIA · Ciberseguridad

Selección del valor de activación en puertas traseras de malware de etiqueta limpia: un análisis de eficacia y detectabilidad

arXiv: 2610.21269v1Tipo de anuncio: nuevo Resumen: el envenenamiento de la puerta trasera de etiqueta limpia es una amenaza para los clasificadores de malware basados en el aprendizaje automático porque las muestras envenenadas se pueden insertar en los datos de entrenamiento sin cambiar sus

La noticia

arXiv:2610.21269v1 Announce Type: new Abstract: Clean-label backdoor poisoning is a threat to machine learning-based malware classifiers because poisoned samples can be inserted into training data without changing their labels. Previous explanation-guided attacks mainly focus on selecting important trigger features, but the values assigned to these features can also affect the attack behavior. In this paper, we study trigger value selection as an independent design factor in clean-label backdoor attacks against feature-based malware classifiers. We evaluate several value-selection strategies on EMBER 2018 and a Win64 PE subset of EMBER 2024 using LightGBM classifiers. We compare rare-value, SHAP-based, frequency-bounded, co-occurrence-aware, and quantile-based selectors under random and distribution-based malware-similar benign sampling. The results