We propose an efficient neural framework for sentence-level discourse analysis in accordance with Rhetorical Structure Theory (RST). Our framework comprises a discourse segmenter to identify the elementary discourse units (EDU) in a text, and a discourse parser that constructs a discourse tree in a top-down fashion. Both the segmenter and the parser are based on Pointer Networks and operate in linear time. Our segmenter yields an F1 score of 95.4, and our parser achieves an F1 score of 81.7 on the aggregated labeled (relation) metric, surpassing previous approaches by a good margin and approaching human agreement on both tasks (98.3 and 83.0 F1)

@inproceedings{lin-et-al-acl-19, abstract = {We propose an efficient neural framework for sentence-level discourse analysis in accordance with Rhetorical Structure Theory (RST). Our framework comprises a discourse segmenter to identify the elementary discourse units (EDU) in a text, and a discourse parser that constructs a discourse tree in a top-down fashion. Both the segmenter and the parser are based on Pointer Networks and operate in linear time. Our segmenter yields an F1 score of 95.4, and our parser achieves an F1 score of 81.7 on the aggregated labeled (relation) metric, surpassing previous approaches by a good margin and approaching human agreement on both tasks (98.3 and 83.0 F1)}, address = {Florence, Italy}, author = {Xiang Lin* and Shafiq Joty* and Prathyusha Jwalapuram and Saiful Bari}, booktitle = {Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics}, numpages = {9}, pages = {4190–4200}, publisher = {ACL}, series = {ACL'19}, title = {A Unified Linear-Time Framework for Sentence-Level Discourse Parsing}, url = {https://www.aclweb.org/anthology/P19-1410/}, year = {2019} }