Session Description:

This session focuses on the emerging frontier of data-driven geological modelling for underground digital twins, aiming to provide reliable representations of subsurface stratigraphy for smart digitalization and practical engineering applications. By integrating site investigation data (e.g., boreholes, CPT, SPT, and geophysical data) with advanced machine learning, stochastic modelling, and uncertainty quantification frameworks, recent studies have demonstrated promising progress toward real-time building and reliable updating of underground geological and geotechnical heterogeneities. Such developments not only improve the understanding of complex ground conditions but are also expected to enhance predictive capability, support risk-informed decision-making, and enable smart infrastructure management across reclamation projects, tunnelling, and excavation, and coastal protection. This session welcomes both methodological innovations and application-driven studies. Topics of interest include, but are not limited to, the following:

• Data-driven geological and geotechnical modelling
• Digital twins for subsurface stratigraphy and underground infrastructures
• Integration of multi-source site investigation data (e.g., boreholes, CPT, SPT, geophysics)
• Machine learning, AI, and large language models for subsurface characterization and digitalization
• Uncertainty quantification and stochastic modelling of geological heterogeneities
• Automated and generative design for geotechnical site planning and smart infrastructure
• Application-driven case studies (e.g., tunnelling, reclamation, excavation, coastal protection)
• Emerging paradigms such as transfer learning, few-shot learning, and immersive technologies (VR/MR) in geological modelling