A machine reading comprehension (MRC) dataset with discourse structure built over multiparty dialog. Molweni's source samples from the Ubuntu Chat Corpus, including 10,000 dialogs comprising 88,303 utterances.
Variants: Molweni
This dataset is used in 2 benchmarks:
| Task | Model | Paper | Date |
|---|---|---|---|
| Discourse Parsing | Structured | Structured Dialogue Discourse Parsing | 2023-06-26 |
| Discourse Parsing | Hierarchical | Improving Multi-Party Dialogue Discourse Parsing … | 2021-10-09 |
| Discourse Parsing | DP | Multi-tasking Dialogue Comprehension with Discourse … | 2021-10-07 |
| Question Answering | Ma et al. - ELECTRA | Enhanced Speaker-aware Multi-party Multi-turn Dialogue … | 2021-09-09 |
| Question Answering | Li and Zhao - BERT | Self- and Pseudo-self-supervised Prediction of … | 2021-09-08 |
| Question Answering | Li and Zhao - ELECTRA | Self- and Pseudo-self-supervised Prediction of … | 2021-09-08 |
| Question Answering | DADgraph | DADgraph: A Discourse-aware Dialogue Graph … | 2021-04-26 |
| Discourse Parsing | Deep Sequential | Molweni: A Challenge Multiparty Dialogues-based … | 2020-04-10 |
Recent papers with results on this dataset: