STORM: Standardized Technology for Optimizing Rapid Modeling
Abstract
This month we have continued to develop the ontology, create software to interface with the various readers and consider how to integrate quantitative models into the CAGs. The creation of the ontology is approached from the top-down by structuring an Upper Model and the bottom-up by populating the Domain Model with specific words and phrases. The multiple reading engines, EIDOS, BBN and SOPHIA produce different output formats requiring software for translation and classification to allow for comparison and merging of CAGs. Strategies for feedback from modeling to reading for improving the reliability of reading are being investigated. Discussions have been initiated to understand how quantitative models and data will be used for creating qualitative model proxies, initializing, parameterizing, and verifying executable models. We also discussed a role for optimization in evaluating executable models. Parallel ontology efforts by different reading teams necessitates intermediate ID matching when incorporating the reading output from multiple teams into the CAGs. There is still a significant amount of manual checking and extracting relations for CAGs and executable models as the automated reading is being advanced. Our plans for next month are to continue creating the ontology using databases such as WordNet and Phrase Net to semi-automatically develop taxonomy terms. Our existing classifier will be developed for improved detection of corroborations, contradictions and extensions. This month we will identify a candidate quantitative model to convert to a qualitative model approximation for integration into our executable modelling framework.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 11, 2018
- Source ID
- W911NF1810017
Entities
People
- Nataša Miškov-Živanov
Organizations
- Army Contracting Command
- Defense Advanced Research Projects Agency
- University of Pittsburgh