In this project, we propose to develop software that employs open source information to assess a community's resilience and preparedness. The goal is challenging because current technology does not scale well due to the heterogeneity of the problem. Specifically, the heterogeneity of the data, as well as the heterogeneity of the assessment process makes it time-consuming to develop extractors for harvesting relevant data, as well as to develop decision methods for performing resiliency/preparedness assessments. Our work in phase I will include identifying detailed use cases and sample data, along with an ontology for the application. We will also develop an end-to-design for a system, OpenWatch, that we will prototype in Phase II. Finally, we will develop machine learning technology to deal with the heterogeneity problem, including algorithms for semi-automatically developing extractors and semi-automatically developing design surfaces for resiliency assessment. The results of the project will include an open-source software architecture for resiliency assessment, upon which a commercial resiliency-assessment service can be built. In addition, aggregated resilience-related data can be repurposed for multiple commercial verticals, including in particular the insurance industry. The technology developed will also contribute to developing a cloud-based service for Web aggregation that can be rapidly customized for new verticals.
