Molweni: A Challenge Multiparty Dialogues-based Machine Reading Comprehension Dataset with Discourse Structure
Li, Jiaqi Liu, Ming Kan, Min-Yen Zheng, Zihao Wang, Zekun Lei, Wenqiang Liu, Ting Qin, Bing
Li, Jiaqi
Liu, Ming
Zheng, Zihao
Wang, Zekun
Liu, Ting
Qin, Bing
Citations
Altmetric:
Alternative Title
Abstract
Research into the area of multiparty dialog has grown considerably over
recent years. We present the Molweni dataset, 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. We annotate 30,066 questions on this corpus, including both
answerable and unanswerable questions. Molweni also uniquely contributes
discourse dependency annotations in a modified Segmented Discourse
Representation Theory (SDRT; Asher et al., 2016) style for all of its
multiparty dialogs, contributing large-scale (78,245 annotated discourse
relations) data to bear on the task of multiparty dialog discourse parsing. Our
experiments show that Molweni is a challenging dataset for current MRC models:
BERT-wwm, a current, strong SQuAD 2.0 performer, achieves only 67.7% F1 on
Molweni's questions, a 20+% significant drop as compared against its SQuAD 2.0
performance.
Keywords
cs.CL, cs.CL
Source Title
Proceedings of the 28th International Conference on Computational Linguistics
Publisher
International Committee on Computational Linguistics
Series/Report No.
Collections
Rights
Date
2020
DOI
10.18653/v1/2020.coling-main.238
Type
Article