Generalized Ontology Discovery Enabling Semantic Search (GODESS) Navy SBIR FY2010.2 Sol No.: Navy SBIR FY2010.2 Topic No.: N102-180 Topic Title: Generalized Ontology Discovery Enabling Semantic Search (GODESS) Proposal No.: N102-180-0708 Firm: Aptima, Inc. 12 Gill Street Suite 1400 Woburn, Massachusetts 01801 Contact: Charlotte Shabarekh Phone: (781) 496-2465 Web Site: aptima.com Abstract: A central challenge in the intelligence community is managing and effectively integrating large amounts of disparate information sources for concise presentation of knowledge to an analyst. Currently, the high volume of near-constant incoming intelligence imposes a substantial burden on the analyst to review and digest the raw intelligence, causing analyst overload and fatigue that could lead to missed intelligence. Generalized Ontology Discovery Enabling Semantic Search (GODESS) addresses this urgent need by analyzing the data's structure and the analyst's information needs to provide customized, data- and goal-driven knowledge extracted from disparate, multi-modal document stores. GODESS automatically discovers systems of related concepts structured in the data to learn ontologies that are optimal for representing the knowledge encapsulated in the database - relearning them as new data sources become available, making the approach tractable for environments with continuous, high-volume information flow. In addition to automatically optimizing semantic search based on the data, GODESS optimizes on the user, returning a dynamic semantic network that targets the user's specific search needs, conditioned on their queries and information objectives. The returned network is a granular graph representation of relevant knowledge, fused at the datastore level, capturing semantic distance between entities providing concise, relevant results across disparate datastores. Benefits: Aptima's GODESS system will provide intelligence analysts with customized semantic search technology allowing them to efficiently process large quantities of multi-modal documents from disparate sources. GODESS dynamically creates structures of related knowledge from the disparate documents and user's search needs, thus customizing the search functionality to best meet the needs of the user and the complexity of the datastore. GODESS reduces an analyst's workload by retrieving relevant entity profiles, concepts and relationships from large, continuously modified datastores, thus eliminating the need to manually comb through original documents or raw output from a text analytics tool. Additionally, GODESS increases an analyst's productivity by anticipating their search needs to uncover knowledge hidden in the data that is intrinsically linked to the search query, therefore providing additional context without the analyst needing to explicitly specify it. GODESS has clear benefits not to just the intelligence community, but also the search engine community, which is striving to improve the relevance of search results and to provide users with compact multi-modal, disparate document knowledge fusion and representations. Return