Transition-based top-down parsing with pointer networks have achieved state-of-the-art results in multiple parsing tasks, while having a linear time complexity. However, the decoder of these parsers has a sequential structure, which does not yield the most appropriate inductive bias for deriving tree structures. In this paper, we propose hierarchical pointer network parsers, and apply them to dependency and discourse parsing tasks. Our results on standard benchmark datasets demonstrate the effectiveness of our approach, outperforming existing methods and setting a new state-of-the-art in both parsing tasks.

@inproceedings{linlin-et-al-emnlp-19, abstract = {Transition-based top-down parsing with pointer networks have achieved state-of-the-art results in multiple parsing tasks, while having a linear time complexity. However, the decoder of these parsers has a sequential structure, which does not yield the most appropriate inductive bias for deriving tree structures. In this paper, we propose hierarchical pointer network parsers, and apply them to dependency and discourse parsing tasks. Our results on standard benchmark datasets demonstrate the effectiveness of our approach, outperforming existing methods and setting a new state-of-the-art in both parsing tasks.}, address = {Hong Kong}, author = {Linlin Liu* and Xiang Lin* and Shafiq Joty and Simeng Han and Lidong Bing}, booktitle = {Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing}, numpages = {9}, pages = {1007–-1017}, publisher = {ACL}, series = {EMNLP'19}, title = {Hierarchical Pointer Net Parsing}, url = {https://www.aclweb.org/anthology/D19-1093.pdf}, year = {2019} }