Interdependent Security and Defense Games: From Airline Security to Vaccination and Cyber Defense

Abstract

Attacks carried out by hackers and terrorists over the last few years have led to increased efforts by both government and the private sector to create and adopt mechanisms to prevent future attacks. This effort has yielded a more focused research attention to models, computational and otherwise, that facilitate and help to improve (homeland) security for both physical infrastructure and cyberspace. In particular, there has been quite a bit of recent research activity in the general area of game-theoretic models for terrorism settings.

The main objective of our project is to design models and algorithms to analyze networked multi-agent system's situations in which each agent is a node/vertex in some network being the source of a potentially deliberate attack by an external malicious agent. The internal network agent is an independent decision-maker assessing the cost-effectiveness of investments in defense or security that eliminates the effectiveness of direct-attack but can only diminish the effectiveness of catching indirect attacks via potential transfer from unprotected neighbors.

We proposed interdependent defense (IDD) games, a computational game-theoretic framework to study aspects of the interdependence of risk and security in multi-agent systems under deliberate external attacks. Our model builds upon Interdependent security (IDS) games. IDS applications include airline and port security, vaccination and cyber defense. IDS games consider the source of the risk to be the result of a fixed randomized-strategy.

We study computational questions in a modified version of the IDS game in which investing in security comes with some level of protection from transfers. More importantly, we adapted IDS games to model the attacker's deliberate behavior. It is important to note that the explicit modeling of risk transfer is an aspect of our model that has not been the focus of previous game-theoretic attacker-defender models of security.

People

Current Participants

• Hau Chan (Ph.D. candidate; Successfully defended his dissertation on June 5th, 2015)
• Luis E. Ortiz

Former Participant

• Michael Ceyko (former undergraduate; went on to M.S. Program in CS at Harvard)

Relevant Publications and Extended Abstracts

• Hau Chan and Luis E. Ortiz. Learning Game Parameters from MSNE: An Application to Learning IDS Games.
  To appear at The 26th International Conference on Game Theory, part of the Stony Brook Game Theory Summer Festival 2015, July 2015.
• Hau Chan and Luis E. Ortiz. Computing Nash Equilibrium in Interdependent Defense Games. In AAAI Conference on Artificial Intelligence, 2015.
  [URL]
• Hau Chan and Luis E. Ortiz. Computing Nash Equilibria in Generalized Interdependent Security Games. In, Advances in Neural Information Processing Systems (NIPS) 27, Z. Ghahramani, M. Welling, C. Cortes, N.D. Lawrence, and K.Q. Weinberger, Editors, Pages 2735--2743, 2014. Curran Associates, Inc.
  [PDF]
  Also presented as invited spotlight and poster at Workshop on Transactional Machine Learning and E-Commerce, Neural Information Processing Systems (NIPS) Dec. 12, 2014, Montreal, Quebec, Canada
• Hau Chan, Michael Ceyko, and Luis E. Ortiz. Interdependent Defense Games: Modeling Interdependent Security under Deliberate Attacks. In Proceedings of the 28th Conference on Uncertainty in Artificial Intelligence (UAI), August 2012.
  Main paper: [PDF], Supplementary material: [PDF]
  A previous related version:
  • Michael Ceyko, Hau Chan, and Luis E. Ortiz. Interdependent Defense Games: Modeling Interdependent Security under Deliberate Attacks (Extended Abstract). In International Conference on Game Theory, 22nd Stony Brook Game Theory Festival of the Game Theory Society, July 2011.
    [PDF]

Software (Prototype)

• Ruby source code for generating random Internet games and solving them using simple smoothed-best-response dynamics

Dataset

• DIMES Autonomous-systems-level Internet network graph for March, 2010

Sponsor