TY - GEN
T1 - Automated Relation Extraction for Improved Generalizability across Different Types of Text
AU - Chen, Qiyang
AU - El-Gohary, Nora
N1 - Publisher Copyright:
© 2024 Computing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation - Selected Papers from the ASCE International Conference on Computing in Civil Engineering 2023. All rights reserved.
PY - 2024
Y1 - 2024
N2 - Bridge textual reports capture technically detailed data/information about bridge conditions and maintenance actions, which offers opportunities to improve the prediction of future bridge conditions for further bridge maintenance decision-making support. To automatically analyze these reports, there is a need for relation extraction methods to extract relation information from the reports for linking recognized entities with predefined semantic categories (e.g., caused by) and representing the extracted semantic relations in a structured way. To address this need, this paper proposes a deep learning-based relation extraction model. The proposed model utilizes convolutional neural network (CNN) to encode sentence-level features, and bidirectional long short-term memory (BiLSTM) to build a relation extractor to capture the patterns of the predefined relation types. The proposed model was evaluated in extracting relations from multiple types of bridge-related textual reports for representing bridge defect information - including relations among bridge entities - in a semantically rich structured way.
AB - Bridge textual reports capture technically detailed data/information about bridge conditions and maintenance actions, which offers opportunities to improve the prediction of future bridge conditions for further bridge maintenance decision-making support. To automatically analyze these reports, there is a need for relation extraction methods to extract relation information from the reports for linking recognized entities with predefined semantic categories (e.g., caused by) and representing the extracted semantic relations in a structured way. To address this need, this paper proposes a deep learning-based relation extraction model. The proposed model utilizes convolutional neural network (CNN) to encode sentence-level features, and bidirectional long short-term memory (BiLSTM) to build a relation extractor to capture the patterns of the predefined relation types. The proposed model was evaluated in extracting relations from multiple types of bridge-related textual reports for representing bridge defect information - including relations among bridge entities - in a semantically rich structured way.
KW - Artificial intelligence
KW - Bridge conditions.
KW - Bridge inspection
KW - Deep learning
KW - Information extraction
KW - Natural language processing
KW - Relation extraction
UR - http://www.scopus.com/inward/record.url?scp=85184281912&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85184281912&partnerID=8YFLogxK
U2 - 10.1061/9780784485231.054
DO - 10.1061/9780784485231.054
M3 - Conference contribution
AN - SCOPUS:85184281912
T3 - Computing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation - Selected Papers from the ASCE International Conference on Computing in Civil Engineering 2023
SP - 451
EP - 458
BT - Computing in Civil Engineering 2023
A2 - Turkan, Yelda
A2 - Louis, Joseph
A2 - Leite, Fernanda
A2 - Ergan, Semiha
PB - American Society of Civil Engineers
T2 - ASCE International Conference on Computing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation, i3CE 2023
Y2 - 25 June 2023 through 28 June 2023
ER -