The IDEAL initiative is developing an innovative information analysis enabling capability that allows users to collaboratively perform example based searches of information to support specific analysis goals. The IDEAL technology leverages KBSI’s powerful ontology-based search solution, JACKALFISH®.
The volume of available information, both human and computer generated, and advances in document management and analysis technologies that provide greater access to data are making it difficult for intelligence analysts and other users to identify and utilize the key information in the data flood. Improvements in semantic modeling, natural language processing, and groupware technology are providing opportunities for addressing these challenges.
KBSI is developing an innovative information analysis enabling capability called Information DiscovEry Assistant that Learns (IDEAL) that will allow users to collaboratively perform example based searches of information in support of specific analysis goals. The IDEAL technologies will enable users to perform searches using a variety of inputs beyond simple hit or miss keyword guesses. In addition, the IDEAL technologies will be able to use these search examples to derive an understanding of analysts’ goals, connecting analysts working on similar problems, or alerting analysts to developments in related areas.
KBSI has extensive experience in semantic search and natural language processing and will be building off of existing KBSI technologies, like the JACKALFISH® search engine, in developing the IDEAL system. The IDEAL technologies will use semantic search capabilities to allow users to search data stores using example texts, knowledge models (ontologies), glossaries and acronym lists. The more input the user provides, the more semantic content can be extracted by the IDEAL query builder, resulting in returns that more closely match the user’s search task. Advanced natural language processing techniques will be used to extract additional information beyond the individual terms in the source and input data and begin building robust semantic indexes rather than simple and limited term indexes for future searches.
The IDEAL technologies will automatically detect similarities among different analysts’ search tasks and use the feedback and results found by one analyst to improve the results of other analysts. In addition, the IDEAL system’s web-based user interface will provide a venue for collaboration by notifying an analyst of other analysts’ related tasks and by sharing the tasks results. This approach, built on enterprise-level technologies, means that IDEAL will be able to scale easily to a corpus of millions of documents and will support hundreds of simultaneous, collaborating users in a web-based environment. An online feedback mechanism will allow users to rate the quality of results returned from each search. This feedback will be processed internally in the IDEAL system and provide the basis for improving the results of future searches.
These capabilities can make IDEAL the foundation for revolutionary improvements in information analysis productivity. In addition to information analysts, any enterprise or web-based search task could benefit greatly from the advanced semantic search, automated collaboration, and feedback-based improvements enabled by the IDEAL technologies. This includes academic and commercial researchers across a broad range of fields, including government and commercial intelligence, security, and biomedical enterprises.