Symbiotic and Integrated Reasoning ENgine (SIREN) for Autonomic Cyber Systems
Abstract
Technical Proposal BAA Number ONRBAA15-001 Symbiotic and Integrated Reasoning ENgine (SIREN) for Autonomic Cyber Systems Technical Contacts: Dr. Guru Prasadh Venkataramani and Dr. Tian Lan The George Washington University, Washington, DC 20052 Email: {guruv, tlan}@gwu.edu, Phone: (202) 994-2980 ONR Program Manager: Dr. J. Sukarno Mertoguno, Division 311 Administrative Contact: Sylvia Ezekilova, The George Washington University Email: osr@gwu.edu, Phone: (202) 994-6255 Period of Performance: 01/01/2015 - 12/31/2017 Project Summary Securing software from external attacks and information leakage is exacerbated by the sophisticated adversaries combined with the increasing complexity of the software modules. In particular, information leakage based attacks and program vulnerability/malfunctions have begun to undermine the reliable use of software. Manual analysis of attack patterns and of system states are insufficient to deal with these everchanging threat scenarios. Software management should be autonomous and have situation awareness such that it can make timely and acurate decisions on counter attack strategies. Closed loop automation techniques usually rely on one of the statistical or formal reasoning strategies for decision making and planning. However, each of them are individually constrained by their inability to deal with uncertainties, explosion in state space or inaccuracy with planning. So, the question that arises is: How to increase the accuracy of decision making and planning while being practical and real time? The solution is motivated by the process of human reasoning, which harnesses both gut-feeling and deliberative thinking. We propose a hybrid approach strategy that synergestically integrates the two methodologies (formal and statistical reasoning) without necessarily unifying the underlying knowledge representations. As we know, formal models are more conducive for planning and statistical methods have better adaptability for uncertainties. The hybrid model derives these advantages from the two constituent approaches by increasing the interactions between them and the partially correlated knowledge representations from the two streams of reasoning models. It allows both the knowledge bases to grow side-by-side and yet interact with each other through close collaboration in the learning and reasoning process. In particular, the interactions between the two sets of knowledge base provides opportunities to bridge the gap when the knowledge in one domain is partial or incomplete (i.e., cross-checking and cross-triggering) and to facilitate cooperative reasoning using both formal and statistical methods. This synergistic solution approach can help both of the knowledge bases to gain richer information over time, and essentially lead to more real time and accurate decision making and planning. Relevance to ONR: ONR has prioritized Cyber Security and Complex Software Systems research as a program, and has designated several thrust areas that emphasize key aspects such as information security and assurance, secure information management and interaction, and determining the security properties of software systems. This proposal seeks to explore more practical and online approaches for autonomous decision making in software systems that guard them against information leakage and violation of integrity. The project will explore novel methodologies to improve the interactions between two existing powerful approaches, namely statistical and formal learning, that can take advantage of their merits while overcoming their demerits. If successful, this project can provide a more robust autonomous planning engine to protect the US Navy’s software and other assets that are controlled by such systems.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2016
- Source ID
- N000141512210
Entities
People
- Guru Prasadh V. Venkataramani
Organizations
- George Washington University
- Office of Naval Research
- United States Navy