Data-Driven Authorship Feature Extraction and Comparative Analysis using Machine Learning

Sonalysts will continue the development of the Authorship and Intent Recognition Environment (AIRE), an adaptive platform that provides the ability to extract authorship and intent from pseudonymously written text. AIRE will use a variety of analytical techniques, including Artificial Intelligence (AI)/Machine Learning (ML) methods, as well as more explainable methods such as goodness of fit and comparative analysis. Sonalysts will continue its user-centered design approach to develop intuitive interfaces and visualizations for Intelligence Analysts. AIRE will leverage AI/ML techniques to find trends and connect the dots among data, and then allow more explainable statistical methods to be run on data to provide more explainability to the Analyst than using “black box” methods alone. These methods will be combined with the use of deep linguistic features such as semantic information and other relevant metadata, to fuse authorship detection and intent recognition methods, rather than treating them as separate routines. This combination of analytical techniques in an intuitive interface will make AIRE a significant innovation and a valuable tool to several end users. Phase II will also represent a significant leap in technical maturity, resulting in the development of a software prototype and integration with representative technologies and networks.