Exploiting Numerical-Contextual Knowledge to Improve Numerical Reasoning in Question Answering
Jeonghwan Kim, Junmo Kang, Kyung-min Kim, Giwon Hong, Sung-Hyon Myaeng
Abstract
Numerical reasoning over text is a challenging subtask in question answering (QA) that requires both the understanding of texts and numbers. However, existing language models in these numerical reasoning QA models tend to overly rely on the pre-existing parametric knowledge at inference time, which commonly causes hallucination in interpreting numbers. Our work proposes a novel attention masked reasoning model, the NC-BERT, that learns to leverage the number-related contextual knowledge to alleviate the over-reliance on parametric knowledge and enhance the numerical reasoning capabilities of the QA model. The empirical results suggest that understanding of numbers in their context by reducing the parametric knowledge influence, and refining numerical information in the number embeddings lead to improved numerical reasoning accuracy and performance in DROP, a numerical QA dataset.
BibTeX
@inproceedings{kim-etal-2022-exploiting,
title = "Exploiting Numerical-Contextual Knowledge to Improve Numerical Reasoning in Question Answering",
author = "Kim, Jeonghwan and
Kang, Junmo and
Kim, Kyung-min and
Hong, Giwon and
Myaeng, Sung-Hyon",
editor = "Carpuat, Marine and
de Marneffe, Marie-Catherine and
Meza Ruiz, Ivan Vladimir",
booktitle = "Findings of the Association for Computational Linguistics: NAACL 2022",
month = jul,
year = "2022",
address = "Seattle, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.findings-naacl.138/",
doi = "10.18653/v1/2022.findings-naacl.138",
pages = "1811--1821"
}