The second official release of OMCSNet-WNLG is now available.

The OMCSNet-WNLG project aims to disambiguate the entire OMCSNet dataset, adding Part-of-Speech, WordNet sense and phrase tagging information to all predicates. This project is also focused on leveraging the semantic relationship information contained within WordNet to further increase the size and accuracy of the OMCSNet-WNLG dataset.

OMCSNet is the largest publicly available dataset of common sense reasoning information. This data is of use to artificial intelligence researchers, natural language processing enthusiasts, and software developers wishing to integrate common sense reasoning into applications.

Information regarding the 1.1 release:

* - Over 30,000 new assertions grafted into the existing OMCSNet dataset, bringing the total to over 315,000.

* - Part-of-Speech tagging for 301,522 words (out of 578,980 total) contained within the OMCSNet assertions.

* - 49,863 OMCSNet concept/data strings (out of 181,395 total) contain at least some WordNet sense tagging.

* - 149,507 OMCSNet concept/data strings (out of 181,395 total) contain at least some Part-of-Speech tagging.

The OMCSNet-WNLG project may be found at:

http://www.eturner.net/omcsnetcpp/wordnet

OMCSNet-WNLG is part of the OMCSNetCPP project. OMCSNetCPP is based on OMCSNet and the Open Mind Common Sense project. Thanks to all individuals associated with these projects for making OMCSNet available.

Note: OMCSNet-WNLG is not officially associated with the Open Mind project. Any questions regarding Open Mind should be referred to http://www.openmind.org/

Regards,

Elliot Turner