More about KIM

This section is for everyone

If you want to try out the KIM Platform and explore its functionalities through the Web UI, see the User's guide (over Latest news).

KIM is a platform for semantic annotation and multi-paradigm search over documents, data, and knowledge.

Off-the-shelf you get extraction of people, locations, organizations, dates, money, and others; a semantic index of your content, and many new ways to search and explore you information space.

KIM comes with:

a pre-loaded ontology and a knowledge base of important entities (people, organizations, ...)
text mining capabilities - to find entities of several types
tools - to upload and annotate new content
user interfaces - to search and navigate your data and content
documentation and training
free (mailing list) and commercial support packages
freedom to change the ontology, text mining algorithms, and user interfaces

KIM is offered in three profiles:

a platform similar in purpose and use to the Knowledge Management (KM) systems ;
a 3rd party extension bundled with larger KM products ( the OEM model );
a SaaS / PaaS, provided remotely with no installation and maintenance efforts.

Semantic annotation in KIM

Semantic annotation is about finding mentions of entities (such as persons, organizations, locations, dates, etc.) in texts. KIM tries to match these mentions to known entities in its Knowledge base (KB). Each entity has a unique Uniform Resource Identifier (URI), properties, description, and aliases, which when matched are attached to the mention. If there is no such entity, KIM automatically generates one with a new URI and description. This process, as well as the result, is called semantic annotation. The automatically generated metadata can later be used for indexing, retrieval, visualization, and hyper-linking of documents.

KIM ontologies

KIM is equipped with an upper-level ontology (PROTON) of about 250 classes and 100 properties, which allows easy bootstrapping of applications. In addition, the platform has a built-in Knowledge Base (KIM KB) , pre-populated with about 200,000 entity descriptions. The idea behind this KB is to provide an exhaustive coverage of all entities of general importance. In this way it forms background knowledge, resembling human common culture. As such, entities that are considered well-known are not typically introduced in documents. Therefore, it is hard to extract their descriptions automatically.

As a technology, the architecture allows all KIM-based applications to perform automatic semantic annotation, content retrieval based on semantic restrictions, as well as querying and modifying the underlying ontologies and knowledge bases.

Search with KIM

In its essence, KIM performs three basic types of search - keyword (full-text search over texts), conceptual (pattern search over the graph of the knowledge base), and combined.

Searching with KIM allows you to:

find all available documents about an event, person, organization, etc.
define search by keywords, entity names, conceptual relations
explore links between entities occurring together in a text
analyze emerging trends about entities of interest over a period of time