A Deep Learning based Approach to Proactive and Reactive Enterprise Security

Abstract: This work presents the construction of a deep learning based algorithmic mechanism to ensure the security of an information system based on proactive and reactive approach. The basic building blocks of the mechanism are threat analytics, cryptographic solutions and adaptive secure multiparty computation. It is basically an attempt of the cross fertilization of algorithmic game theory and financial cryptography. It defines deep learning strategy for proactive and reactive security based on the properties of adaptive secure multi-party computation. This work also shows the application of the mechanism on reasoning two test cases: (a) enterprise security and (b) malicious learning system in adversarial setting. It analyzes the complexity of the mechanism in terms of computational cost and security intelligence. This study is expected to be useful for a defense research organization. Keywords: Adaptive Secure multi-party computation, Financial cryptography, Proactive security, Reactive security, Deep learning, Enterprise security, Adversarial machine learning, Complexity analysis. This work is organized as follows. Section 2 highlights the importance of adaptive secure multi-party computation in deep learning mechanism. It defines deep learning strategy of information system security based on proactive and reactive approaches. Section 3 outlines the algorithmic mechanism of deep learning in details in the context of analyzing two test cases: DLM-ES for enterprise security and DLM–MLS for adversarial learning system. Section 4 shows the complexity analysis of deep learning mechanism from the perspective of computational cost and security intelligence. Section 5 concludes the work.