PREDICTING PROTEST: DEMONSTRATING THE VALUE AND LIMITS OF NEURAL NETWORK MODELS TO ENHANCE MILITARY DECISION MAKING
Loading...
Authors
Schill, Michael T.
Subjects
artificial intelligence
AI
neural networks
transformer architecture
social media analysis
predictive models
natural language processing
NLP
military decision making
protests and riots
multilingual language models
multi-label binary classification
AI applications
AI
neural networks
transformer architecture
social media analysis
predictive models
natural language processing
NLP
military decision making
protests and riots
multilingual language models
multi-label binary classification
AI applications
Advisors
Warren, Timothy C.
Date of Issue
2024-12
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
This research demonstrates the utility of an artificial intelligence (AI) model to predict protests and riots near locations of interest. This work enhances the U.S. Air Force’s Agile Combat Employment (ACE) concept by providing early warnings of civil unrest near airfield locations, creating decision advantage for commanders. There is little research into using neural network transformer architecture in this way, to predict social behavior at specific geolocations. To advance the field, I engineered a full AI model training pipeline and data post-processing to produce an interactive dashboard for military users. Multilingual social media posts from 2016 emanating from 56 countries were used to fine-tune an AI model to predict protests or riots near an airport in the next fifteen days. The model correctly predicts 85% of future protests or riots, with a false positive rate of 35%. This research shows the value of AI neural networks for prediction of protests and riots and highlights important caveats for military decision makers to consider when implementing such a model.
Type
Thesis
Description
Series/Report No
Department
Organization
Identifiers
NPS Report Number
Sponsors
Funder
Format
Citation
Distribution Statement
Distribution Statement A. Approved for public release: Distribution is unlimited.
Rights
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.