Large-scale language technology resources exist for syntactic analysis (the Penn Treebank for English, Negra and Tiger for German) and for shallow semantic analysis (thematic roles, word senses, anaphora, time expressions). There are also some attempts to bring these together (OntoNotes). But large integrated annotated resources for deep semantic analysis do not exist yet. We argue that such a resource could cause a breakthrough in computational semantics and show the first steps in designing one based on Discourse Representation Theory.

If you would like to meet with the speaker, please contact Stefan Thater.