BERT-based Spatial Information Extraction

Hyeong Jin Shin, Jeong Yeon Park, Dae Bum Yuk, Jae Sung Lee


Abstract
Spatial information extraction is essential to understand geographical information in text. This task is largely divided to two subtasks: spatial element extraction and spatial relation extraction. In this paper, we utilize BERT (Devlin et al., 2018), which is very effective for many natural language processing applications. We propose a BERT-based spatial information extraction model, which uses BERT for spatial element extraction and R-BERT (Wu and He, 2019) for spatial relation extraction. The model was evaluated with the SemEval 2015 dataset. The result showed a 15.4% point increase in spatial element extraction and an 8.2% point increase in spatial relation extraction in comparison to the baseline model (Nichols and Botros, 2015).
Anthology ID:
2020.splu-1.2
Volume:
Proceedings of the Third International Workshop on Spatial Language Understanding
Month:
November
Year:
2020
Address:
Online
Editors:
Parisa Kordjamshidi, Archna Bhatia, Malihe Alikhani, Jason Baldridge, Mohit Bansal, Marie-Francine Moens
Venue:
SpLU
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
10–17
Language:
URL:
https://aclanthology.org/2020.splu-1.2
DOI:
10.18653/v1/2020.splu-1.2
Bibkey:
Cite (ACL):
Hyeong Jin Shin, Jeong Yeon Park, Dae Bum Yuk, and Jae Sung Lee. 2020. BERT-based Spatial Information Extraction. In Proceedings of the Third International Workshop on Spatial Language Understanding, pages 10–17, Online. Association for Computational Linguistics.
Cite (Informal):
BERT-based Spatial Information Extraction (Shin et al., SpLU 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.splu-1.2.pdf
Video:
 https://slideslive.com/38940078