Cypher 1.9 Demo

[Sherman Monroe]

Hi All,

I'm pleased to announce the Cypher 1.9 demo available now online. This announcement proceeds the official release announcement to come in the next few days containing feature details. The demo runs the Monrai Standard dataset, and can be pointed against several SPARQL Endpoints, including dbpedia.

Cypher is a natural language transcoder which generates the RDF triple or SPARQL query representation of phrases and sentences.

Some example outputs are:

-  the person who played in a film (in dbpedia) 
-  the writers who played in a film (in dbpedia) 
-  the actors wrote what (in dbpedia) 
-  the actors who are married wrote what (in dbpedia) 
-  the presidents who were married to someone wrote what (in dbpedia)
-  the people who wrote some books (in dbpedia) 
-  Bruce Willis stars in what (in dbpedia) 
-  the people who were influenced by actors (in dbpedia)

Click the links in the (typically first) result set to view the raw generated SPARQL, and the second icon to see the result set (having problems with the vsual browser, should be fixed tommorrow). The biggest remaining bottle-neck is the lexicon. For each parsed input, there must exists a lexical entry which is used to construct the triples. For example, the lexemes actor, writer, and star all specify RDF templates for dbp:starring property. A more robust lexicon for Dbpedia (via UMBEL and Cyc) is underway. In addition, there is a collaborative-database of lexical data that is in the works to help aleviate this shortcoming and cover more ontologies and public datasets.

Please join the Cypher Mailing list to keep up with release updates, etc.

Enjoy,
-sherman