Last month, I began the discussion of an emerging generation of new knowledge management technology that can unleash the power of this knowledge. The power of this knowledge can become available through methods like automatic semantic tagging, user profiling, knowledge query transforms, and knowledge hunting through the powerful concept of knowledge recontextualization (a really big word implying that you're able to change the context of the knowledge).

In this second article, we continue our discussion of technology components in the new generation of knowledge management technology. In a later article, we'll discuss typical applications, the new and powerful concept of active-reflexive data, and iconographic drag and drop user-customizable component-based application assembly. So stay tuned to future issues where the fun will continue!

Automatic classification
Results context can be automatically organized as semantic pivots around common themes that reflect the common context of the results collection, bound by the common themes across participating kSigs in kThreads. The themes in the results kThread provide an automatic taxonomy of the results and can be mapped to any corporate standard taxonomy through a taxonomy connector.

Furthermore, corporate taxonomy mapping can be automatically learned by the system through intelligent mediation, contextualizing documents that are already indexed to the targeted taxonomy. The binding medium becomes the kThreads that bind the mediation document kSigs to their respective taxonomy indexes, automatically generating the cross-association references.

In fact, such a system could automatically maintain the taxonomy-mapped semantic mediation fabric by conforming to a corporate taxonomy mapping policy, which mandates that taxonomy mapping must preserve a specified level of confidence before mapped intersections apply. This becomes straightforward to achieve in an autonomous fashion using the above pattern matching metaphor.

Document knowledge views
Classification taxonomies can be published showing documents and their participating paragraphs componentized around themes along the kThread in which they participate.