EMNLP Findings2023
GLGR: Question-aware Global-to-Local Graph Reasoning for Multi-party Dialogue Reading Comprehension
Yanling Li, Bowei Zou, Yifan Fan, Xibo Li, Ai Ti Aw, Yu Hong
Abstract
Graph reasoning contributes to the integration of discretely-distributed attentive information (clues) for Multi-party Dialogue Reading Comprehension (MDRC). This is attributed primarily to multi-hop reasoning over global conversational structures. However, existing approaches barely apply questions for anti-noise graph reasoning. More seriously, the local semantic structures in utterances are neglected, although they are beneficial for bridging across semantically-related clues. In this paper, we propose a question-aware global-to-local graph reasoning approach. It expands the canonical Interlocutor-Utterance graph by introducing a question node, enabling comprehensive global graph reasoning. More importantly, it constructs a semantic-role graph for each utterance, and accordingly performs local graph reasoning conditioned on the semantic relations. We design a two-stage encoder network to implement the progressive reasoning from the global graph to local. The experiments on the benchmark datasets Molweni and FriendsQA show that our approach yields significant improvements, compared to BERT and ELECTRA baselines. It achieves 73.6% and 77.2% F1-scores on Molweni and FriendsQA, respectively, outperforming state-of-the-art methods that employ different pretrained language models as backbones.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
1 top-tier paper has cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- Structural Characterization for Dialogue Disentanglement cites ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
- Back to the Future: Bidirectional Information Decoupling Network for Multi-turn Dialogue Modeling cites ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
- Back to the Future: Bidirectional Information Decoupling Network for Multi-turn Dialogue Modeling cites Self- and Pseudo-self-supervised Prediction of Speaker and Key-utterance for Multi-party Dialogue Reading Comprehension
- Self- and Pseudo-self-supervised Prediction of Speaker and Key-utterance for Multi-party Dialogue Reading Comprehension cites ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
Cited by top-tier papers1
Builds on5
- ELECTRA: Pre-training Text Encoders as Discriminators Rather Than GeneratorsKevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. ManningICLR 2020 · 541 citations
- Self- and Pseudo-self-supervised Prediction of Speaker and Key-utterance for Multi-party Dialogue Reading ComprehensionYiyang Li, Hai ZhaoEMNLP Findings 2021 · 27 citations
- SRLGRN: Semantic Role Labeling Graph Reasoning NetworkChen Zheng, Parisa KordjamshidiEMNLP 2020 · 22 citations
- Structural Characterization for Dialogue DisentanglementXinbei Ma, Zhuosheng Zhang, Hai ZhaoACL 2022 · 20 citations
- Back to the Future: Bidirectional Information Decoupling Network for Multi-turn Dialogue ModelingYiyang Li, Hai Zhao, Zhuosheng ZhangEMNLP 2022 · 9 citations
Related papers
- Multimodal Dialog System: Relational Graph-based Context-aware Question UnderstandingHaoyu Zhang, Meng Liu, Zan Gao, Xiaoqiang Lei et al.ACM MM 2021 · 31 citations
- Question-Interlocutor Scope Realized Graph Modeling over Key Utterances for Dialogue Reading ComprehensionJiangnan Li, Mo Yu, Fandong Meng, Zheng Lin et al.ACL Findings 2023 · 3 citations
- Fact-Driven Logical Reasoning for Machine Reading ComprehensionSiru Ouyang, Zhuosheng Zhang, Hai ZhaoAAAI 2024 · 10 citations
- Recurrent Chunking Mechanisms for Long-Text Machine Reading ComprehensionHongyu Gong, Yelong Shen, Dian Yu, Jianshu Chen et al.ACL 2020 · 39 citations
- Exploiting Reasoning Chains for Multi-hop Science Question AnsweringWeiwen Xu, Yang Deng, Huihui Zhang, Deng Cai et al.EMNLP Findings 2021 · 16 citations
