The government has a wealth of policy knowledge derived from military specialists, regulators, and subject matter experts (SMEs) in myriad fields. What it lacked, until now, was a flexible tool for searching the content of the thousands of policies using the knowledge of those experts. The LMI Research Institute funded development of a tool—OpenPolicy™—to provide agencies with the ability to capture the knowledge of their experts and use it to intuitively search their massive storehouse of policy.
“Anyone who uses the Internet understands the limitations of traditional search engines. You end up with document-level results,” explained Gus Creedon, program manager with LMI. “There’s no simple way to search the content of policy documents and drill down to the appropriate paragraph.”
Deep and Intelligent Searches
OpenPolicy™ solves this problem. The tool, which took a six-person team two years to develop and cost $1 million, combines Ontotext U.S.A.’s semantic database platform, LMI SME-developed search ontologies, and computing power to scan the text of tens of thousands pages. Based on domain-specific ontologies, the tool also suggests possible search terms and phrases to help users refine their requests and obtain better results.
For agencies wanting to use OpenPolicy™, LMI initially builds a powerful computing environment to host the knowledgebase. It then loads all of an agency’s documents—policies, regulations, meeting notes, trouble tickets, essentially any text-based file—into the database. The system can scale to store billions of individual paragraphs.
Reusable Ontologies Capture Expertise
To help users mine the policy knowledgebase, LMI SMEs develop unique ontologies for each policy area. These ontologies capture SME knowledge and include key terms, synonyms, industry slang, related concepts, and business processes. For example, users wanting to search on policies related to ammunition would also find ammo, ordnance, munitions, explosives, fired brass, 40 mm machine gun, and many other terms on the search menu. As a result, users can search tens of thousands of pages across hundreds of individual documents and quickly be directed to specific paragraphs within those documents.
Semantic Searches Drive Benefits
Both the Office of the Secretary of Defense, Logistics and Material Readiness, and the Department of Health and Human Services are using OpenPolicy™ to query large policy knowledge bases, some as large as 15,000 to 20,000 pages of text. In addition to speeding queries and producing more usable search results, the tool offers significant benefits, including:
- Capturing SME knowledge via the reusable ontologies.
- Improving policy governance and quality.
- Accelerating research and training.
- Increasing compliance.
Development of OpenPolicy™ was a natural fit for LMI. “We’ve been investigating and experimenting with the technology for nearly a decade,” noted Creedon. “Plus, we have subject matter experts across all areas of government policy.”
Gus Creedon is a senior consultant in LMI’s Digital Services service line. With LMI for 30 years, he is responsible for the design, development, and delivery of software products, including databases and web-based systems, cloud migrations for multi-tier systems, and blockchain system engineering solutions.