Identifying relevant common sense information in knowledge graphs

Guy Aglionby, Simone Teufel


Abstract
Knowledge graphs are often used to store common sense information that is useful for various tasks. However, the extraction of contextually-relevant knowledge is an unsolved problem, and current approaches are relatively simple. Here we introduce a triple selection method based on a ranking model and find that it improves question answering accuracy over existing methods. We additionally investigate methods to ensure that extracted triples form a connected graph. Graph connectivity is important for model interpretability, as paths are frequently used as explanations for the reasoning that connects question and answer.
Anthology ID:
2022.csrr-1.1
Volume:
Proceedings of the First Workshop on Commonsense Representation and Reasoning (CSRR 2022)
Month:
May
Year:
2022
Address:
Dublin, Ireland
Venue:
CSRR
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1–7
Language:
URL:
https://aclanthology.org/2022.csrr-1.1
DOI:
10.18653/v1/2022.csrr-1.1
Bibkey:
Cite (ACL):
Guy Aglionby and Simone Teufel. 2022. Identifying relevant common sense information in knowledge graphs. In Proceedings of the First Workshop on Commonsense Representation and Reasoning (CSRR 2022), pages 1–7, Dublin, Ireland. Association for Computational Linguistics.
Cite (Informal):
Identifying relevant common sense information in knowledge graphs (Aglionby & Teufel, CSRR 2022)
Copy Citation:
PDF:
https://preview.aclanthology.org/auto-file-uploads/2022.csrr-1.1.pdf
Video:
 https://preview.aclanthology.org/auto-file-uploads/2022.csrr-1.1.mp4
Code
 guyaglionby/kg-common-sense-extraction
Data
CommonsenseQAConceptNetOpenBookQA