Beyond development: Challenges in deploying machine learning models for structural engineering applications

TL;DR

This paper discusses the challenges of deploying machine learning models in structural engineering, emphasizing issues like overfitting, data representativeness, and validation techniques to improve real-world application readiness.

Contribution

It provides practical insights and illustrative examples highlighting key challenges and solutions for deploying ML models in structural engineering contexts.

Findings

Model overfitting and underspecification are critical issues.

Rigorous validation and adaptive sampling improve model deployment.

Physics-informed feature selection enhances model generalizability.

Abstract

Machine learning (ML)-based solutions are rapidly changing the landscape of many fields, including structural engineering. Despite their promising performance, these approaches are usually only demonstrated as proof-of-concept in structural engineering, and are rarely deployed for real-world applications. This paper aims to illustrate the challenges of developing ML models suitable for deployment through two illustrative examples. Among various pitfalls, the presented discussion focuses on model overfitting and underspecification, training data representativeness, variable omission bias, and cross-validation. The results highlight the importance of implementing rigorous model validation techniques through adaptive sampling, careful physics-informed feature selection, and considerations of both model complexity and generalizability.