End-to-End Discourse Parsing on RST-DT (test)
92.02Span ScoreConditional Splitting Framework
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Conditional Splitting Frameworksegmentation setting=end-to-end2021.06 | 92.02 | 87.05 | 78.82 | |
| Lin et al. (2019)training strategy=joint2021.06 | 91.75 | 86.38 | 77.52 | |
| Lin et al. (2019)training strategy=pipeline2021.06 | 91.14 | 85.8 | 76.94 | |
| Joty et al. (2012)2021.06 | 82.4 | 76.6 | 67.5 | |
| Soricut and Marcu (2003)2021.06 | 76.7 | 70.2 | 58 |