Discourse relation is crucial for Natural Language Processing, and it has
received a lot of research attention especially after the release of PBTB v2.0. Within the domain on discourse relation, implicit discourse relation recognition is important since it takes up more than half of all discourse relations. At the same time, it is a difficult task, and the performance of state-of-the-art
systems is far from making them viable for real world applications.
This project aims to leverage on both traditional feature-based and deep learning approaches to improve the recognition performance of PDTB style implicit discourse relation.
Proceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI'17), 2017.