2014 Data (Information)
MASC Word Sense Sentence Corpus, Crowdsourced subset
The MASC Word Sense Sentence corpus, Crowdsourced subset, is distributed as a set of three *tsv files (tab-separated format) that contain the sentences, annotation labels, and WordNet senses of the corpus. For 45 of the 116 words used from the original MASC Word Sense Sentence corpus (http://dx.doi.org/10.7916/D80V89XH), there are up to 1000 sentences per word drawn from the heterogeneous MASC corpus, with sense labels from WordNet. Each sentence exemplifies at least one MASC word, annotated for its WordNet sense. Each word/sentence pair has up to 25 crowdsourced sense labels collected on Amazon Mechanical Turk.
- MASC_AMT_RELEASE_1.0.tar.gz application/x-gzip 6.89 MB Download File
More About This Work
- Academic Units
- Computer Science
- Published Here
- June 27, 2014
This is a dataset on word sense annotation with a comprehensive README file that describes the contents.