<![CDATA[Bridges are critical infrastructure highly vulnerable to earthquake-induced damage, posing serious risks to transportation continuity and public safety. Traditional methods, such as visual inspection, remain in use; however, they are limited in efficiency, scalability, and accuracy. This highlights the urgent need for more advanced approaches. Machine Learning (ML) and Artificial Intelligence (AI) have emerged as promising alternatives for assessing post-earthquake bridge damage. However, existing studies often lack a systematic synthesis of methodological trends, rely on limited or unstandardized datasets, and insufficiently address real-world implementation challenges. This study conducts a Systematic Literature Review (SLR) to critically examine the application of ML/AI in post-earthquake bridge damage assessment, focusing on methodological trends, commonly used datasets, and implementation challenges. Relevant journal articles published between 2019 and 2025 were selected through structured keyword strategies and filtered based on publication type, relevance, and journal quality (Q1-Q4). The findings indicate that Random Forest (RF) and Convolutional Neural Networks (CNN) are among the most widely applied ML methods, owing to their strengths in classification and visual data analysis. Frequently used datasets include bridge damage records from California, shake table test time-series data, and sensor-based monitoring data. Persistent challenges include data heterogeneity, limited availability of real-time datasets, and the interpretability of ML models. The novelty of this study lies in providing a consolidated synthesis of current research, bridging methodological gaps, and highlighting implementation challenges. Future research should focus on developing real-time datasets, establishing robust model validation frameworks, and enhancing the interpretability of ML techniques to strengthen disaster risk mitigation and improve bridge resilience.]]>
Copyrights © 2025
