Statistical Foundations for Measurement-Based System Verification in Complex Networks

Abstract

Complex networks of various types are central to the current and future mission readiness of the Air Force. Particularly important are communication networks, which can span multiple temporal and spatial scales and connect a variety of human, semi-autonomous, and autonomous agents, each playing an important role towards the overall mission success. In order to preserve such network-based information infrastructures (and, more generally, the structure and integrity of the information itself traversing these networks), questions of their design and robustness are critical. But upon deployment, a corresponding paradigm of measurement and system verification becomes equally if not more critical. And with this need for measurement in complex mobile communication networks comes the need for an appropriate statistical methodology, for everything from sampling, to characterization, to modeling, inference, and prediction of network-oriented parameters associated with performance and, ultimately, risk. However, statistics for many of the most basic and fundamental tasks, while developed to an arguably mature state in now-canonical contexts like signal and image processing, data mining, etc., are still comparatively immature in the context of complex networks. In fact, tools for even some of the seemingly most innocuous of statistical tasks, such as equipping summary statistics (e.g., density, clustering coefficients, etc.) of measured networks with confidence intervals, are almost non-existent for complex networks. Accordingly, under this award we pursued a broad-based, multi-faceted program of research to systematically lay key pieces of the statistical foundation necessary to pursue measurement-based systems validation in complex networks. This included work on estimation under network sampling, propagation of uncertainty in network summary statistics, robust dynamic community detection, network topology inference, and network averaging under repeated sampling.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 02, 2019
Accession Number
AD1096383

Entities

People

  • Eric D. Kolaczyk

Organizations

  • Boston University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Air Force
  • Big Data
  • Communication Networks
  • Computational Science
  • Data Analysis
  • Data Mining
  • Data Science
  • Information Processing
  • Information Science
  • Measurement
  • Network Science
  • Networks
  • Signal Processing
  • Social Media
  • Statistical Analysis
  • Statistical Inference
  • Statistics

Fields of Study

  • Computer science

Readers

  • Computer Networking
  • Regression Analysis.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference