Semantical Machine Understanding
Navy SBIR 2006.2 - Topic N06-153 ONR - Ms. Cathy Nodgaard - [email protected] Opens: June 14, 2006 - Closes: July 14, 2006 N06-153 TITLE: Semantical Machine Understanding TECHNOLOGY AREAS: Information Systems, Human Systems OBJECTIVE: Develop a quick-reaction capability to extract specified semantic content from large volumes of unstructured multilingual text and represent it mathematically. Use these representations to identify events, relationships and trends to enable interoperable knowledge sharing and intelligence analysis across joint and coalition forces. DESCRIPTION: Defense Transformation has changed warfighting tactics from use of platform-based large-scale initiatives to quick reaction, team-based mobile force operations in discrete events. There will be increased operations with Joint, Coalition, Non-Government and Volunteer Organizations which will require analysis of open-source (uncertain, conflicting, partial, non-official) data. Teams will consist of culturally diverse partners with rapidly changing team members and various organizational structures. These characteristics will put increasingly difficult demands on short turn-around, high stakes, crisis driven intelligence analysis. In order to respond to this challenge, more powerful information analysis tools are needed that can quickly extract meaning and intent from large volumes of data. There are a number of extant tools for data mining including advanced search engines, key word analysis and tagging technology but better tools are needed to achieve advanced information discovery which provide more focused and directed content rather than line-item search results. Key to such a capability is automated understanding of intent or meaning and the ability to represent it in a language/ culture free format. This solicitation proposes building such a capability by selecting the most promising data search and visualization/presentation tools and combining them with newly developing semantic analysis tools for the delivery of event-specific information. The final product will be a data/text analysis tool that can perform semantic searches and provide language/culture-free information in a format that can be used for discovery of events, relationships and trends. PHASE I: Review current technology and tools for semantic search methods that have established empirical evidence of effectiveness. Select top candidate/s, or combinations thereof for matching with commercially available text analysis tools for development of a prototype concept of machine understanding of meaning. Incorporate these tools into a computational model or automated agent to enable the collection of real-time data which can provide the capability to search by entity, concept or meaning extraction. PHASE II: Develop a web-based experimental testbed of the search technology in the domain of intelligence analysis for coalitions operations. Develop a prototype, based on empirically validated techniques and evaluate in a simulated or representative operational environment. Provide metrics and measures to assess performance improvement in an intelligence analysis or decision making environment. Provide functional and design guidelines to assure the extensibility of the prototype to other operational venues. PHASE III: Incorporate the search technology in a planned operational test environment at a Navy or Defense Intelligence Analysis venue. Validate, standardize and document underlying software for application purposes and implement in a field experiment. Coordinate with user subject matter experts to instantiate a working model with actual data, get user commitment for training and maintenance of the application. Collect performance data to validate improved performance Develop guidelines and documentation for tool transition to an operational setting. PRIVATE SECTOR COMMERCIAL POTENTIAL/DUAL-USE APPLICATIONS: Private-sector applications would include any information analysis situation that involves high data volume and quick response requirements. This would include state and local emergency support teams for crisis action planning and humanitarian aid response. Business applications, such as corporate knowledge management or textual research would also be a target. REFERENCES: 2. Gerber, Cheryl (2005) Smart Searching. In Military Information Technology, Vol 9, Issue 9. 3. Foltz, P. W. (2002) Quantitative cognitive models of Text and Discourse Processing. In The Handbook of Discourse Processes. Mahwah, NJ: Lawrence Erlbaum Publishing. KEYWORDS: Semantics, text mining, machine learning, discourse analysis, knowledge management TPOC: Mike Letsky
|