Information Network Embedding for Scalable and Self Adaptive Analysis of Terrorist Organizations
Abstract
Terrorism is an ongoing, persistent threat to U.S. national security, with active terrorist groups such as Al-Quaeda and ISIS carrying out attacks worldwide. Terrorist network analysis is critical for predicting, preventing, and countering terrorist attacks because the terrorist network represents all of the information available about terrorists, their organizations, and actions. However, network analysis is challenging because terrorist organizations and their modus operandi are continuously changing. For each change, the existing decisions systems together with the predictive models that inform them may become ineffective requiring expensive and time-consuming expert intervention to retune the models and decision systems. To counter and pre-emptively respond to the network dynamics, this project aims to conduct preliminary research to define an information network embedding that is able to preserve network topological properties relevant for counter-terrorism applications. Once the embedding procedure is obtained, the produced features can be used to perform terrorist network analysis and develop self-adaptive prediction models. To be able to test the preservation of the structural properties by the resulting embedding, we will perform a preliminary data collection from open-sources such as Wikipedia and Wikidata to build an initial terrorist information network.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 06, 2019
- Source ID
- W911NF1910438
Entities
People
- Edoardo Serra
Organizations
- Army Contracting Command
- Boise State University
- United States Army