Efficient training framework for deep learning–based structural response prediction under strong ground motions

doi:10.1016/j.engstruct.2026.123073

Efficient training framework for deep learning–based structural response prediction under strong ground motions

2026-06-11

0

PRE

AI

DOI：10.1016/j.engstruct.2026.123073

Original

Original request for help

Save

Abstract

En 中文

• Autoencoder-based framework for efficient seismic response prediction. • Latent-space maximum-diversity selection of strong ground motions. • Improved generalization using sparse nonlinear dynamic analyses. • Validated accuracy and robustness via numerical case studies.

Efficient training framework for deep learning–based structural response prediction under strong ground motions

Abstract

Journal

Organization