TY - GEN
T1 - EviDR
T2 - 10th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2021
AU - Zhou, Yongwei
AU - Bao, Junwei
AU - Sun, Haipeng
AU - Liang, Jiahui
AU - Wu, Youzheng
AU - He, Xiaodong
AU - Zhou, Bowen
AU - Zhao, Tiejun
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021
Y1 - 2021
N2 - Reasoning machine reading comprehension (R-MRC) aims to answer complex questions that require discrete reasoning based on text. To support discrete reasoning, evidence, typically the concise textual fragments that describe question-related facts, including topic entities and attribute values, are crucial clues from question to answer. However, previous end-to-end methods that achieve state-of-the-art performance rarely solve the problem by paying enough emphasis on the modeling of evidence, missing the opportunity to further improve the model’s reasoning ability for R-MRC. To alleviate the above issue, in this paper, we propose an Evidence-emphasized Discrete Reasoning approach (EviDR), in which sentence and clause level evidence is first detected based on distant supervision, and then used to drive a reasoning module implemented with a relational heterogeneous graph convolutional network to derive answers. Extensive experiments are conducted on DROP (discrete reasoning over paragraphs) dataset, and the results demonstrate the effectiveness of our proposed approach. In addition, qualitative analysis verifies the capability of the proposed evidence-emphasized discrete reasoning for R-MRC (Code is released at https://github.com/JD-AI-Research-NLP/EviDR ).
AB - Reasoning machine reading comprehension (R-MRC) aims to answer complex questions that require discrete reasoning based on text. To support discrete reasoning, evidence, typically the concise textual fragments that describe question-related facts, including topic entities and attribute values, are crucial clues from question to answer. However, previous end-to-end methods that achieve state-of-the-art performance rarely solve the problem by paying enough emphasis on the modeling of evidence, missing the opportunity to further improve the model’s reasoning ability for R-MRC. To alleviate the above issue, in this paper, we propose an Evidence-emphasized Discrete Reasoning approach (EviDR), in which sentence and clause level evidence is first detected based on distant supervision, and then used to drive a reasoning module implemented with a relational heterogeneous graph convolutional network to derive answers. Extensive experiments are conducted on DROP (discrete reasoning over paragraphs) dataset, and the results demonstrate the effectiveness of our proposed approach. In addition, qualitative analysis verifies the capability of the proposed evidence-emphasized discrete reasoning for R-MRC (Code is released at https://github.com/JD-AI-Research-NLP/EviDR ).
KW - Discrete reasoning
KW - Evidence
KW - Machine reading comprehension
UR - https://www.scopus.com/pages/publications/85118105481
U2 - 10.1007/978-3-030-88480-2_35
DO - 10.1007/978-3-030-88480-2_35
M3 - 会议稿件
AN - SCOPUS:85118105481
SN - 9783030884796
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 439
EP - 452
BT - Natural Language Processing and Chinese Computing - 10th CCF International Conference, NLPCC 2021, Proceedings
A2 - Wang, Lu
A2 - Feng, Yansong
A2 - Hong, Yu
A2 - He, Ruifang
PB - Springer Science and Business Media Deutschland GmbH
Y2 - 13 October 2021 through 17 October 2021
ER -