Session Description:

In soil mechanics and geotechnical engineering, traditional physics-based methods require large domain knowledge and tend to take a large amount of time, labour and costs to set up models and calibrate model parameters. Conversely, pure data-driven approaches require vast amounts of high-quality data and tend to behave unpredictably outside the training domain. Currently, integrating existing knowledge and data-driven methods has emerged as a potential method to leverage the merits of both methodologies and has been applied in many domains. In this context, this mini-symposium (MS) aims to explore and discuss the state-of-the-art development and applications of knowledge-informed data-driven methods in geotechnics and risk assessment. The scope of this MS includes, but is not limited to, the following topics:

• Theory-guided, scientific, or knowledge-informed machine learning algorithms
• Knowledge-informed data-driven methods for solving BVPs and inverse problems
• Knowledge-informed data-driven constitutive modelling
• Hybrid FEM and data-driven modelling
• Data-efficient modelling integrating high-cost and low-cost data
• Site characterisation and spatial variability modelling
• Uncertainty quantification and reliability-based design