Reference Number: ARL-SRP-00027
Project Description
Modern cyber-physical systems operate in environments where adversarial interactions span interconnected domains, including cyber networks and physical processes. Actions in one domain influence observability, capabilities, and outcomes in others, creating complex strategic dependencies. Deception plays a central role, not only through explicit mechanisms such as decoys or honeypots, but also through strategic behavior that shapes an adversary’s beliefs and decisions over time.
This project proposes a foundational study of multi-domain deception through the lens of game theory, and learning. The goal is to develop principled models and scalable solution methods for designing and evaluating strategies in adversarial environments with partial and imperfect information. We aim to develop a general framework that captures belief manipulation, cross-domain interactions, and adaptive strategic behavior.
The potential research thrusts are: (1) modeling deception as belief shaping under incomplete information games, (2) integrating learning-based methods for adaptive and data-driven deception, and (3) developing scalable algorithms for large, structured multi-domain systems.
Technical Skills
- Artificial Intelligence/Machine Learning
- Communication and Networking Technologies
- Computer Modeling/Simulation
- Computer Science
- Cybersecurity
- Data Analysis
- Data Science
- Data Visualization
- Mathematics
- Natural Language Processing
- Statistics
- Software Development
- Game Theory
