Featured research

01Ongoing

Korea Occupational Safety and Health Agency

A Study on the Effectiveness Analysis and Improvement of On-Site Operability of Hazard and Risk Prevention Plans in the Construction Industry.

Publications

Figure 5. A confusion matrix of the GPT-3.0 fine-tuned model for (a) both combined structured and unstructured data and (b) unstructured data alone.

Applied Sciences·2024

Harnessing Generative Pre-Trained Transformers for Construction Accident Prediction with Saliency Visualization

Byunghee Yoo, Jinwoo Kim, Seongeun Park, Changbum R. Ahn, Taekeun Oh
A fine-tuned GPT model predicts six construction accident types from accident-report text and highlights words that influence its predictions.
construction safetyaccident predictionnatural language processinglarge language modelsGPTtext classificationexplainable AIsaliency visualization
0View publication