Agency Deploys Knowledge Management Solution to Improve Search Capabilities
Every day, analysts from the Office of Naval Intelligence (ONI) pore through thousands of separate pieces of data, from intercepted phone calls to cryptic messages on websites.
Intelligence analysts must piece that information together into timely, effective intelligence, a task compounded by data streams that may contain different spellings, meanings and associations. For instance, the Library of Congress has detected 32 different ways to spell the name "Muammar Khaddafi".
This one example illustrates the massive challenge facing ONI analysts. How do they sort through multiple threads of data to identify potentially ominous trends and create plausible, cohesive analyses that will provide actionable intelligence to the appropriate law enforcement or military units?
ONI chose to deploy topic maps to help them organize snippets of information semantically to find multiple associations or relationships among sets of data. These topic maps can help ONI analysts retrieve the precise information they are seeking, and thus prepare more timely, detailed and effective reports.
ONI turned to Innodata Isogen on the basis of our internal expertise and experience building the information architecture required to support dynamic flexible topic maps. We have used our extensive information science expertise to create customized topic maps for a diverse set of clients.
Innodata Isogen's depth of experience with sophisticated information management and classification tools, such as XML mark-up languages, played a key role, enabling the project team to set-up a structured approach for ONI. Ultimately, the team’s aim was to show ONI analysts how to use the customized topic map tools to organize their data into taxonomies, ontologies and schemata, to help them link and retrieve relevant information.
Designing topic maps for an entity that processes millions of bits of data per day is hardly an easy task. The sensitive nature of the content added to the complexity. Because Innodata Isogen consultants have extensive clearance to access the data, the team was able to teach ONI analysts how to create their own taxonomies, so they could continue to populate the topic maps with the correct terms, phrases and associations.
As a first step, the project team began building an interim authoring solution tool that would allow the ONI analysts to create the categories that would eventually be used to build the topic maps and ontologies. At the same time, ONI began developing business process and usecase examples to provide the parameters Innodata Isogen would use to decide how information would be gathered and assembled. Innodata Isogen next built mock-ups of the user interface, which were shared with ONI analysts in multiple demonstrations to give them an opportunity to suggest revisions. This ongoing dialogue served as a critical checkpoint to ensure that the solution met the client’s needs.
One of the most frustrating revelations from the 9/11 investigations was that various law enforcement agencies had obtained information about one or more Al Qaeda terrorists, but were unable to use that information effectively or share it with other agencies in time to respond. If they had been using more effective search technologies, however, the intelligence analysts would have been able to search through multiple databases and find new associations within the data. When the ONI project is finished, analysts will be able to aggregate information from the various databases, retrieve data more efficiently and provide more relevant information to other agencies.
Innodata Isogen is now working with ONI to ensure that the relationships between the various classifications of data are accurate. They are also building subject-based classification systems to look at research requests in context and leave analysts additional room for interpretation.
Ultimately, ONI analysts will be able to read the summary of a document and then decide whether to read more or move on to another, so that actually opening the file becomes unnecessary. This more precise search feature will also allow them to exclude an entire range of topics and focus on the exact associations they are looking for within a particular subject. This will help them analyze more information at a faster rate and deliver better intelligence.