Éva Mújdricza-Maydt

Data

Semantic Annotation:

• SemAnno Corpus: combined verb sense and semantic role annotation: for appr. 3500 annotated verbal predicates, the predicates are annotated with GermaNet 9.0 senses, and their arguments are annotated for each of the predicate senses with VerbNet-style semantic roles.
• SR3de Corpus: parallel dataset for German SRL with PropBank, VerbNet, and FrameNet annotations of appr. 3000 predicate argument structures out of the CoNLL 2009 shared task German data.

Sample data for the CRFalign aligner:

• OpenSubtitles (315 German-English file pairs)
• Gutenberg (114 German-English file pairs)