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

TL;DR

EviDR enhances reasoning machine reading comprehension by emphasizing evidence detection and using a graph convolutional network to improve discrete reasoning, demonstrating effectiveness on the DROP dataset.

Contribution

The paper introduces an evidence-emphasized discrete reasoning framework that detects evidence and employs a graph network to improve R-MRC performance.

Findings

Significant performance improvement on DROP dataset

Effective evidence detection at sentence and clause levels

Qualitative analysis confirms reasoning capability

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.