Postdoctoral fellow position in Deep Machine Learning and Probabilistic Modelling for the Inversion of Mechanical Engineering Problems
Deep Machine Learning and Probabilistic Modelling for the Inversion of Mechanical Engineering Problems:
The project is focussed on electromagnetic (EM) measurements that incorporate information about the resistivity distribution of the Earth´s subsurface. Such information can be used to determine the porosity of the rocks and the type of fluids contained within those rocks. The correct interpretation (numerical inversion) of the measurements is critical for obtaining an accurate map of the Earth´s subsurface.
The main objective of this Project is to implement and analyse the new Bayesian inversion methodologies proposed in BCAM for the efficient inversion of geophysical EM resistivity measurements. Such methodologies rely on advanced importance sampling techniques and adaptive numerical schemes. The possibility to apply the new methods in machine learning algorithms will be also investigated.