Abstract, Rationale, Stance: A Joint Model for Scientific Claim Verification

TL;DR

This paper introduces ARSJoint, a joint model for scientific claim verification that improves accuracy by simultaneously learning abstract retrieval, rationale selection, and stance prediction, reducing error propagation.

Contribution

The paper proposes a novel joint learning framework for scientific claim verification that integrates multiple tasks and enhances information sharing among them.

Findings

Outperforms existing methods on SciFact dataset

Effectively reduces error propagation among tasks

Improves accuracy by joint task learning

Abstract

Scientific claim verification can help the researchers to easily find the target scientific papers with the sentence evidence from a large corpus for the given claim. Some existing works propose pipeline models on the three tasks of abstract retrieval, rationale selection and stance prediction. Such works have the problems of error propagation among the modules in the pipeline and lack of sharing valuable information among modules. We thus propose an approach, named as ARSJoint, that jointly learns the modules for the three tasks with a machine reading comprehension framework by including claim information. In addition, we enhance the information exchanges and constraints among tasks by proposing a regularization term between the sentence attention scores of abstract retrieval and the estimated outputs of rational selection. The experimental results on the benchmark dataset SciFact show that our approach outperforms the existing works.