Adaptive Support Attack Detection Algorithm for Stacked Autoencoder Based on Adversarial Learning

Collaborative Filtering (CF) recommendation systems are pivotal in modern online platforms, providing users with personalized content and enhancing user experience. However, these systems are vulnerable to various attacks that can manipulate recommendations. This paper introduces an innovative approach to securing recommendation systems by implementing an adaptive Stacked Autoencoder detection algorithm based on adversarial learning. We propose a method that leverages the power of adversarial networks to identify and mitigate attacks in real time while maintaining high-quality recommendations for legitimate users. Our algorithm adapts to evolving attack strategies, providing robust protection against malicious activities. Experimental results demonstrate the effectiveness of our approach in safeguarding recommendation systems from adversarial manipulation.