Skip to main navigation Skip to search Skip to main content

EviDR: Evidence-Emphasized Discrete Reasoning for Reasoning Machine Reading Comprehension

  • Yongwei Zhou
  • , Junwei Bao
  • , Haipeng Sun
  • , Jiahui Liang
  • , Youzheng Wu
  • , Xiaodong He
  • , Bowen Zhou
  • , Tiejun Zhao*
  • *Corresponding author for this work
  • Harbin Institute of Technology
  • JD AI Research

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

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 ).

Original languageEnglish
Title of host publicationNatural Language Processing and Chinese Computing - 10th CCF International Conference, NLPCC 2021, Proceedings
EditorsLu Wang, Yansong Feng, Yu Hong, Ruifang He
PublisherSpringer Science and Business Media Deutschland GmbH
Pages439-452
Number of pages14
ISBN (Print)9783030884796
DOIs
StatePublished - 2021
Event10th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2021 - Qingdao, China
Duration: 13 Oct 202117 Oct 2021

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13028 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference10th CCF Conference on Natural Language Processing and Chinese Computing, NLPCC 2021
Country/TerritoryChina
CityQingdao
Period13/10/2117/10/21

Keywords

  • Discrete reasoning
  • Evidence
  • Machine reading comprehension

Fingerprint

Dive into the research topics of 'EviDR: Evidence-Emphasized Discrete Reasoning for Reasoning Machine Reading Comprehension'. Together they form a unique fingerprint.

Cite this