Artificial Intelligence, Machine Learning, and Deep Learning in Structural Engineering: A Scientometrics Review of Trends and Best Practices

Artificial Intelligence (AI), machine learning (ML), and deep learning (DL) are emerging techniques capable of delivering elegant and affordable solutions which can surpass those obtained through traditional methods. Despite the recent and rapid advancements in developing next-gen AI-based techniques, we continue to lack a systemic understanding of how AI, ML, and DL can fundamentally be integrated into the structural engineering domain. To advocate for a smooth and expedite the adoption of AI techniques into our field, we present a state-of-the-art review that is specifically tailored to structural engineers. This review aims to serve three purposes: (1) introduce the art and science of AI, ML, and DL in terms of its commonly used algorithms and techniques with particular attention to those of high value to this domain, (2) map the current knowledge within this domain through a scientometrics analysis of more than 4000 scholarly works with a focus on those published in the last decade to identify best practices in terms of procedures, performance metrics, and dataset size etc., and (3) review past and recent efforts that applied AI derivatives into the various subfields within structural engineering. Special attention is given to the application of AI, ML, and DL in earthquake, wind, and fire engineering, as well as structural health monitoring, damage detection, and prediction of properties of structural materials as collected from over 200 sources. Finally, a discussion on trends, recommendations, best practices, and advanced topics towards the end of this review.