• Eleni Koronaki
• Anina Šarkić Glumac, University of Luxembourg
• Stéphane P.A. Bordas

The advances in computer-aided design and computational methods have contributed to the increased understanding and enhanced modelling of Real World and Industrial applications. Moreover, rapid changes are continiously underway due to advances in data analytics, AI, machine learning, the internet of things, etc. It can surely be said that advent of these methods has presented a new analytical paradigm in a variety of science and engineering applications.

Nevertheless, in Real World or Industrial applications it is not always pragmatic to assemble the data necessary for the implementation of data-hungry deep learning methods. In addition, more often than not, data are available from different modalities, i.e. different instrumentation and/or low as well as high-fidelity simulations. Furthermore, the opaqueness of "black-box"-type machine-learning frameworks, does not allow for increased physical insight or enhanced understanding of the underlying system/process that is under investigation.

These short-comings notwithstanding, data-driven algorithms, are attractive for Industrial and Real World problems because they combine, favourably, speed and accuracy versus the traditional modelling approaches like Computational Fluid Dynamics (CFD).

Contributions to this mini-symposium are invited to address these unique challenges linked to tackling Real world and industrial applications including, but not limited to:
- Data assimilation
- Hybrid equation-based and data-driven modelling
- Reduced order modelling, multi-fidelity surrogate modeling
- Machine learning, deep learning
- Data-driven modelling as a means of protecting industrial/process information
- Uncertainty quantification
- Performance-based design
- Digital twins