This minisymposium focuses on uncertainty quantification for engineering systems, a domain where the computational expense of high-fidelity simulations poses significant challenges. Traditional methods as standard Monte Carlo simulation, Polynomial Chaos and Gaussian Process Regression often become prohibitively costly. This necessitates the development and application of innovative and efficient methodologies. A particular emphasis of this minisymposium is the advancement of surrogate models. Surrogates approximate the input-output relation or other functional dependencies within a model using inexpensive, easy-to-evaluate functions. They are derived from the underlying governing equations or from a very limited number of high-fidelity simulations. Consequently, they allow to reduce the computational burden for uncertainty quantification. This minisymposium aims to showcase the latest advancements in surrogate modelling for computationally expensive engineering simulations. In addition, the application to certification, reliability analysis and parameter identification is of interest. Possible topics for contributions of the minisymposium are: - Advances in surrogate modelling - Polynomial-based surrogates - Advanced Gaussian process regressions - Operator-Learning - Multi-fidelity surrogates - Adaptive surrogates and active learning strategies