Defense Against National Vulnerabilities in Public Data
Abstract
The evolving diversity of new data sources brings with it a host of unrevealed vulnerabilities that could have significant impacts on national security. The initial goal of the project was to examine how these emerging data sources posed risks to exposing domestic national security interests through unintentional disclosure of sensitive information through open data like social media. IST Research architected and prototyped a proof-of-concept system that can provide automatic feedback on the measurable risk inherent with various collections of data. The designed platform provides algorithmic computation in real time, not only generating results that are machine-readable but are structured to be understandable by humans. This back-end provides the stream of potential vulnerabilities based on algorithmic orchestration geared towards specific subject matter domains. The platform is a lightweight distributed platform that allows for the quick orchestration and training of algorithms from a variety of domains and programming languages. The philosophy used provides a simple and efficient platform for identifying, sequencing and training a custom set of algorithms and data geared to a specific domain and problem set. Specifically, this is an approach and platform designed from its inception to ingest multi-dimensional real-time data and provide a dynamic metadata service to link the most appropriate algorithms with all incoming data sources.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 28, 2017
- Accession Number
- AD1033364
Entities
People
- Rich Salaway