Automated Relation Extraction for Improved Generalizability across Different Types of Text

Qiyang Chen, Nora El-Gohary

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

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.

Original languageEnglish (US)
Title of host publicationComputing in Civil Engineering 2023
Subtitle of host publicationVisualization, Information Modeling, and Simulation - Selected Papers from the ASCE International Conference on Computing in Civil Engineering 2023
EditorsYelda Turkan, Joseph Louis, Fernanda Leite, Semiha Ergan
PublisherAmerican Society of Civil Engineers
Pages451-458
Number of pages8
ISBN (Electronic)9780784485231
DOIs
StatePublished - 2024
Externally publishedYes
EventASCE International Conference on Computing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation, i3CE 2023 - Corvallis, United States
Duration: Jun 25 2023Jun 28 2023

Publication series

NameComputing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation - Selected Papers from the ASCE International Conference on Computing in Civil Engineering 2023

Conference

ConferenceASCE International Conference on Computing in Civil Engineering 2023: Visualization, Information Modeling, and Simulation, i3CE 2023
Country/TerritoryUnited States
CityCorvallis
Period6/25/236/28/23

Keywords

  • Artificial intelligence
  • Bridge conditions.
  • Bridge inspection
  • Deep learning
  • Information extraction
  • Natural language processing
  • Relation extraction

ASJC Scopus subject areas

  • General Computer Science
  • Civil and Structural Engineering

Fingerprint

Dive into the research topics of 'Automated Relation Extraction for Improved Generalizability across Different Types of Text'. Together they form a unique fingerprint.

Cite this