DIMACS Working Group on Applications of Order Theory to Homeland Defense and Computer Security

September 30, 2004
DIMACS Center, CoRE Building, Rutgers University

Jonathan Farley, Massachusetts Institute of Technology
Anthony A. Harkin, Harvard University, harkin@deas.harvard.edu
Mel Janowitz, DIMACS / Rutgers University, melj@dimacs.rutgers.edu
Stefan Schmidt, Physical Science Laboratory, schmidt@psl.nmsu.edu
DIMACS Workshop on Applications of Order Theory to Homeland Defense and Computer Security.


Jason Crampton, University of London

Title: Using Cryptographic Techniques to Enforce an Information Flow Policy

Information flow policies can be used to articulate confidentiality requirements in highly sensitive commercial and military computer systems. An information flow policy assumes the existence of a partially ordered set of (security) labels and a function that assigns a label to users and data items. The policy requires that all information flow should respect the ordering on the labels: in particular, if a user reads a data object (causing information to flow from the object to the user), then the user's label should be greater than or equal to the object's label. A variety of cryptographic methods have been proposed for implementing this "no read down" aspect of an information flow policy. We will describe some topics of current research and ideas for future research in this area.

Vladimir A. Lefebvre, University of California, Irvine

Title: Modeling a Moral choice

A real human choice is predetermined by moral factors in addition to the utilitarian ones. Unlike the majority of the existing theories of choice based on the ideas of utility, the reflexive theory allows us to include moral factors into models of choice. With this approach, alternatives are characterized not only by their utility, but also by their "purity." The new method recently elaborated allows us to compute the preferences of the alternatives using their utility-measures and purity, based on experts' evaluations.

Stefan Schmidt, Physical Science Laboratory, New Mexico State University &
Phoenix Mathematical Systems Modeling, Inc.

Title: An Information-Theoretic Approach to Manipulation

We establish a connection between manipulation and noisy coding. In order to accomplish this, we introduce the notion of a manipulation channel. The goal of successful manipulation is to find a sequence of actions that best controls the sequence of reactions being related to the actions via a manipulation channel. Within our setting, we compare our proposed measures for the degree of manipulability and predictability. Finally, we will show how our approach might help to improve homeland security.

Stefan E. Schmidt, Physical Science Laboratory, New Mexico State University &
Phoenix Mathematical Systems Modeling, Inc.

Title: Modeling and Simulating Terrorist Recruitment via Reflexive Theory

The challenge in modeling terrorist recruitment activity lies in the difficulty of factoring in the human psyche. Common modeling approaches understand human decision making processes primarily as the optimization of utility gain based only on rationality. However, a disregard of the moral dimension within terrorist behavior makes it impossible adequately to describe this situation. For example, a suicide mission requires a high degree of willingness for self-sacrifice, which goes beyond the concept of a homo economicus. Reflexive Theory allows modeling the combination of external, environmental and internal, moral factors. Therefore, in this joint work, Tim Kaiser and I have chosen the reflexive approach to formalize terrorist recruitment processes. In contrast to utilitarian optimization, the principle of reflexivity proposes that the subject attempts to reach a state of congruence between the self and the internal model of the self.

Previous: Program
Workshop Index
DIMACS Homepage
Contacting the Center
Document last modified on September 14, 2004.