UMBEL is designed to help content interoperate on the Web
UMBEL is a general reference structure of 28,000 concepts, which provides a scaffolding to link and interoperate other datasets and domain vocabularies
UMBEL is a base vocabulary for the construction of other concept-based domain ontologies, also designed for interoperation
UMBEL is accessible as set of free Web Service endpoints and as a Clojure/Java API
UMBEL's reference concepts form a coherent graph of relationships, organized into 33 mostly disjoint SuperTypes. Closely aligned concepts are clustered into hierarchically organized classes with further relatedness links. Each concept in the knowledge graph may act as a facet. Logical operations – including inferencing – may be conducted amongst concepts. UMBEL is based on open W3C standards. UMBEL is a curated subset of OpenCyc, which leads as well to Cyc and many additional, useful reasoning tools.
Each of UMBEL's 28,000 concepts is a "thing, not a string". Each concept is defined and is characterized by a semset, comprised of a rich set of possible synonyms, aliases, acronyms and jargon. The canonical URI for each concept provides a persistent identifier. Labels may be expressed in any language, making UMBEL inherently multi-lingual. Combined with related concepts, these semantic characterizations enable the correct meaning to be chosen for a given context.
Inferencing and the use of broader semsets means that searches are not limited to keywords or labels. The essence of concepts may now be searched, including related neighbors in the knowledge graph. These neighbor concepts and their semsets may be further leveraged to expand queries, thus enabling searching within targeted sections of the graph. Try UMBEL's powerful search function for yourself.
The UMBEL vocabulary is designed for mapping ontologies and external vocabularies. The vocabulary provides expressions of likelihood relationships distinct from exact identity or equivalence. UMBEL has about 48,000 formal mappings to OpenCyc, DBpedia, PROTON, GeoNames, and schema.org, and provides linkages to more than 2 million Wikipedia entities (English version).
Specific concepts or entire sections of the UMBEL knowledge graph provide tie-in points for relating concepts in external schema, vocabularies or domain ontologies. UMBEL's concepts thus act as shared reference points across information domains, helping to ensure that context and orientation are maintained as multiple datasets or sources are brought together. UMBEL's reference concepts provide a fixed structure for bridging differences in external information sources.
A key portion of the UMBEL knowledge graph models attributes, used to describe and characterize entities, things and concepts. UMBEL attributes thus provide a means for relating actual instance data across various external sources. Because of its use of the W3C's OWL 2 language standard, both concepts (schema) and instances (data and their characterizations) can interoperate in reference to UMBEL.
Unlike gold standards for instance data such as Wikipedia or Wikidata (among others), UMBEL is not intended to be a source of entity content. Nor is UMBEL intended to be a replacement for domain schema or ontologies. Rather, as a coherent and comprehensive knowledge scaffolding, UMBEL is designed to be a gold-standard reference for conceptual and attribute interoperability on the Web.