Inferring and Hiding Structural Properties of Complex Networks

Abstract

Complex networks often encode sensitive information. For example, the topology and attributes of a computer network can reveal important information about the network operator’s capabilities and critical infrastructure. Due to this inherent sensitivity, it is often important to infer or hide properties of a given network, such as network’s topology, or the evolution of a dynamic processes over the network. In practice, both inference and hiding can be challenging due to a combination of incomplete information and sensitivity to graph topology and random process dynamics. These challenges are exacerbated by evolving application domains and adversarial capabilities. In this proposal, we will design and analyze algorithms for inferring and hiding properties of complex networks based on emerging applications and adversarial models. We consider two main problems of interest, each of which will constitute a thrust of this proposal: (1) Identifying and hiding the source or destination of a message routed over a graph. (2) Identifying and hiding the topology of a graph based on partial observations. Although these are old problems, work in this space has predominantly focused on simple graph and observation models that do not apply to today’s complex networks; this disconnect calls for new algorithms, fundamental limits, and analysis techniques.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 21, 2022
Source ID
FA95502110090XX0

Entities

People

  • Giulia Fanti

Organizations

  • Air Force Office of Scientific Research
  • Carnegie Mellon University
  • United States Air Force

Tags

Fields of Study

  • Computer science

Readers

  • Computer Networking
  • Educational Psychology
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • Space