Monoidal computers, networks and strategic learning: Methods for Adaptive Defense in Cyber Security (MADCybS)

Abstract

This project offers entirely new semantics to capture adversarial threats and their moving-target defensive strategies in terms of the so-called monoidal computers. This new method allows for compositional reasoning about composed systems as well as analyzing the computational and logical complexity of defensive strategies, and the hardness of cyber attacks. To model strategic adaptation for defenders against cyber attacks, the PI formulates a new concept of "learning equilibrium" that circumvents the restricted assumptions associated with the Nash-equilibrium traditional framework. Moreover, the PI will investigate the problem of how a player may be able to predict his/her opponents strategies before they can predict him/her.

Document Details

Document Type
DoD Grant Award
Publication Date
Mar 23, 2016
Source ID
FA95501510263

Entities

People

  • Dusko Pavlovic

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of HawaiĘ»i System

Tags

Readers

  • Cybersecurity.
  • Game Theory.
  • Neural Network Machine Learning.

Technology Areas

  • Cyber
  • Cyber - Cryptography