Application of advanced machine learning in civil engineering: A survey

Machine learning (ML) has been increasingly adopted across various disciplines, including civil engineering (CE), to address a wide range of complex problems. This study conducts a systematic literature review to examine recent trends in the ML applications within CE and to identify key challenges associated with its implementation. The review is proposed focusing on four research questions concerning data scarcity, efficient construction of learning datasets, overfitting mitigation, and the integration of CE's multidisciplinary nature. The analysis focuses on five major fields in CE— structural, geotechnical, transportation, water and environmental, and energy engineering—and evaluates the application of five prominent ML techniques: multilayer perceptron, convolutional neural network, recurrent neural network, generative adversarial network, and reinforcement learning. A total of 800 ML studies in CE were reviewed. Key subfields within each CE domain were identified, and domain-specific applications of ML were synthesized to address the predefined research questions. The findings of this study provide practical insights and methodological guidance for researchers aiming to apply ML to real-world CE challenges in a robust and informed manner.