Enhancing Time-Series Access Control Using Deep Recurrent Neural Networks and Generative Adversarial Networks

Abstract

In this research, we introduce an innovative Attribute-Based Access Control (ABAC) system incorporating the novel attribute of "access history" and deep learning techniques, specifically time-series neural networks such as Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) networks, to enhance decision-making accuracy. The system includes key components such as Policy Decision and Enforcement Points, access log management, and an offline-trained deep learning model for real-time access validation. Our proposed method improves accuracy by integrating Generative Adversarial Networks (GANs) to generate realistic synthetic data for training. Experimental results on various datasets, including real-world and synthetic data, demonstrate the superior performance of our model over traditional and recent methods, achieving an accuracy of over 98% in complex access control scenarios.