Explainable Resiliency Graphs
Modeling, Analyzing, and Prioritizing Explainable Defenses for Cyber-Physical System Resiliency.
The adoption of digital technology in industrial control systems (ICS) enables improved control over operation, ease of system diagnostics and reduction in cost of maintenance of cyber physical systems (CPS). However, digital systems expose CPS to cyber-attacks. The problem is grave since these cyber-attacks can lead to cascading failures affecting safety in CPS. Unfortunately, the relationship between safety events and cyber-attacks in ICS is ill-understood and how cyber-attacks can lead to cascading failures affecting safety. Consequently, operators are ill-prepared to handle cyber-attacks on CPS. This project will develop a formal framework and associated toolsets based on Machine Learning and AI Planning that will enable operators to analyze how a cyber-attack can trigger safety events in CPS and/or cascading failures and then interactively determine potential approaches to mitigate those threats. The formal framework, which we call an Explainable Resiliency Graph, is based on the notion of transition system and is built through a composition of attack trees representing cyber-attacks in the CPS and fault trees representing failures. The transition system would be represented as an AI Planning problem in a new language called Resiliency Graph Description Language (RGDL) that we develop exclusively for this problem domain. RGDL is an extension of the classical Planning Domain Definition Language (PDDL) and leverages all its power for solving AI planning problems. The reasoning engine is based on AI Planning and would explore plans that identify various attack events in the CPS that can lead to cascading failures and safety issues. The accompanying tools would allow the operator to interact with the AI planner to query the underlying transition system and perform what-if analysis. This analysis would provide actionable suggestions from the tool including insights into potential attack vectors and help prioritize efforts to secure critical assets effectively. These suggestions would comprise of a diverse set of solutions each of which can potentially take the CPS to a safe and secure state. To help the operator decide which of the actionable suggestions to implement, the toolset would provide explanations in natural language.
Resiliency of Mission-Critical Systems
Mission-critical systems are exposed to significant risks from faults and cyberattacks. To effectively navigate these challenges, it is essential to create missions that can anticipate, withstand, recover from, and adapt to disruptions. In this project, we develop a framework that facilitates the design of cyber-resilient missions. This is achieved by defining a mission as a workflow, transforming it into a formal Coloured Petri Net (CPN) model, deriving threat models, and generating CPN-based attack representations. By analyzing the state transitions of the mission under various attack scenarios, the process can determine whether missions succeed, fail, or remain incomplete. This approach is particularly useful during the early stages of analysis and can help inform mission restrictions that enhance resiliency.
Project’s Latest News
Publications & Software
- Rakesh Podder, Turgay Caglar, Shadaab Bashir, Sarath Sreedharan, Indrajit Ray, and Indrakshi Ray, “SPEAR: Security Posture Evaluation using AI Planner-Reasoning on Attack-Connectivity Hypergraphs”, In Proceedings of the 30th ACM Symposium on Access Control Models and Technologies (SACMAT), Stony Brook, NY, USA, July 2025.
- Shadaab Bashir, Rakesh Podder, Sarath Sreedharan, Indrakshi Ray, and Indrajit Ray, “Resiliency Graphs: Modelling the Interplay between Cyber Attacks and System Failures through AI Planning”, In Proceedings of the 6th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA), Atlanta, GA, USA, October 2024. Paper
- Indrajit Ray, Sarath Sreedharan, Rakesh Podder, Shadaab Bashir, and Indrakshi Ray, “Explainable AI for Prioritizing and Deploying Defenses for Cyber-Physical System Resiliency”, In Proceedings of the 5th IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA), Atlanta, GA, USA, October 2023. Paper
- Mahmoud Abdelgawad and Indrakshi Ray, “Resiliency Analysis of Mission-critical System of Systems Using Formal Methods”, In Proceedings of the 38th IFIP Annual Conference on Data and Applications Security and Privacy (DBSec), San Jose, CA, USA, July 2024. Paper
- Mahmoud Abdelgawad and Indrakshi Ray, “Methodology for resiliency analysis of mission-critical systems”, In Proceedings of the 39th ACM/SIGAPP Symposium on Applied Computing, Avila, Spain, April 2024. Paper
- Mahmoud Abdelgawad, Indrakshi Ray, and Tomas Vasquez, “Workflow resilience for mission-critical systems”, In Proceedings of the International Symposium on Stabilizing, Safety, and Security of Distributed Systems (SSS), New Jersey, USA, October 2023. Paper
Our Team
Faculty Members
Students
Our Sponsers
