Credible Data-driven Models

The group aims at developing, implementing and analyzing models and methods accounting for their credibility, and assimilating data into the model. The credibility concept embraces three underlying ideas: control of the numerical accuracy (Verification), monitor the pertinence of the model (Validation) and account for the aleatoric nature of the systems analyzed (Uncertainty Quantification). The data assimilation strategies are incorporating into the models the information contained in data from sensors, observations and other models. This is complementary to the Validation phase (via parameter identification) and strongly related with the Uncertainty Quantification.

The group has a proven track record in applying these tools and methods to diverse disciplines in applied sciences and engineering. In particular, some of the current active projects and research lines pertain to the field of Automotive Engineering, Geophysical Modelling, and Biomechanics (bio metamaterials).