Kommentar |
We will briefly overview different kind of target structures and existing ressources (discourse treebanks, graphbanks, dependency annotated corpora, relation sets). We will discuss generic inference tasks and subtasks of interest (sentence/text-level, labeled/unlabeled discourse parsing, segmentation, attachement, inference of explicit/implicit relations), and a range of different computational methods one can apply to these (e.g., bottom-up parsers with distinct decision methods and linguistic features, structured learning approaches, discriminative reranking, (recursive) neural nets). |