This Small Business Innovation Research (SBIR) Phase II project improves access to knowledge by auto-organizing unstructured data to respond to specific individuals, groups, and their activities. Taxonomize Resource Aid (TRA) uses syntactic, semiotic, semantic and statistical techniques to generate and update resource taxonomies, which are multi-level indices into the information corpus (documents, web sites, etc.) specific to users' activities. These active taxonomies are practice-relevant and personalized, and they provide applications of enhanced search, auto-produced portals, personalized content management, and knowledge discovery. For example, TRA's coordinated knowledge directories produce discovery of trends in time-based documents (such as discussion groups); extraction of information from unstructured data (such as distributed themes); and notifications from monitoring multiple information sources for patterns of confluences (e.g., news relevant to collaborating partners) or discrepancies (e.g., knowledge missing in one area that can be filled from another). Phase II development will take the successful prototype that was tested in educational settings, and create a commercial product (initially as a SOAP/WSDL web service) that will be licensed to firms selling software solutions in the areas of e-learning, search, and knowledge management.

Taxonomize Resource Aid (TRA) will provide knowledge tailored for individuals, groups, and activities, and thus will provide people who have been limited by accessibility, resources, or background ready access to resources of knowledge, instruction, and collaboration. The TRA prototype has already been shown to provide significant benefits to some university students who were learning how to do primary research. Those who have difficulty with the culture, language, or technology gain the greatest benefits from TRA, because it gives them accelerated access to knowledge that is automatically selected for relevance to their activities, based on Taxonomize's powerful auto-categorization capabilities. TRA can help in any field where people need to organize, manage, access and use large amounts of information and resources. TRA can help improve education, healthcare, defense, and government organizations process information quickly, especially when dealing with immediate and critical situations. It can also help disadvantaged people find necessary resources, and keep updated with changes that would otherwise be infeasible to monitor. TRA improves knowledge accessibility, flexibility and adaptability and affordability of general learning capabilities, and so may benefit formal and informal learning in every area.

Project Start
Project End
Budget Start
2004-08-01
Budget End
2006-07-31
Support Year
Fiscal Year
2004
Total Cost
$511,956
Indirect Cost
Name
Taxonomize
Department
Type
DUNS #
City
Cupertino
State
CA
Country
United States
Zip Code
95014