RAG-Enhanced Spam Detection Framework: A Synergy of Cloud Scalability and Adaptive AI Models

RAG-Enhanced Spam Detection Framework: A Synergy of Cloud Scalability and Adaptive AI Models

연구 분야: Safety

학회: SN Computer Science

As more communications increasingly depend on email, spam identification becomes the focus of contemporary digital systems. In this paper, a better real-time spam detection architecture that incorporates Retrieval-Augmented Generation (RAG) within a scalable cloud-based system is proposed. The new framework is designed to merge sophisticated machine learning models such as Random Forest, Long Short-Term Memory (LSTM), and Transformer-based models with RAG to better capture context and adaptive decision-making. By employing AWS services like EC2, S3, and Kubernetes, the system facilitates real-time large-scale email processing with cost-effectiveness. RAG facilitates dynamic knowledge retrieval, which greatly enhances the identification of new and complex spam patterns. Empirical assessments show that the system with RAG achieves high classification accuracy (up to 99.2%) and low false positives over traditional standalone ML methods. Security measures like AES-256 encryp- tion, JWT-based API token authentication, and regular vulnerability scanning through OWASP ZAP and SonarCloud additionally support the system’s solidity. Upcoming improvements involve integration with high-frequency threat feeds and user-controlled filtering according to client policies and sophisticated spam con- ditions. This paper presents a sound and responsive platform for next-generation spam protection systems that are scalable, secure, and sensitive to emerging dangers in digital communication.