Hamed Saidaoui

Hamed Saidaoui Copyright: © Martin Braun
Lehrstuhl für Mathematics for Uncertainty Quantification


Room: 159

Pontdriesch 14-16

52062 Aachen


Phone: +49 15739694614


Hamed is a postdoctoral scientist in RWTH working in the field of physics-informed neural networks. Prior to joining RWTH, Hamed worked at Qatar Foundation (Qatar) for about two and half years in the “Qatar Environment and Energy Research Institute" as a postdoctoral researcher. His research was mainly focused on applying Machine Learning techniques to solve physical problems. Hamed obtained his Ph.D. from KAUST in 2016 in Computational Physics with his thesis entitled “ Impact of disorder on spin-dependent transport”.


Research Focus

Hamed’s main research consists of developing efficient machine learning models to solve complicated and demanding partial differential equations(PDEs). The latter are the main building blocks of most scientific and engineering problems. Nevertheless, solving them witnesses many difficulties ranging from the lack of accuracy of classical methods to the curse of dimensionality which renders it impossible to tackle PDEs in higher dimensions. Machine learning has been proven to overcome the curse of dimensionality issues with its special sampling techniques, meanwhile lacking enough accuracy to compete with classical solvers and sometimes being dependent on the use of very deep and costly neural network architectures. His work consists of improving the predictions of machine learning-based solvers in terms of accuracy and complication.


Research Areas

Applying Machine learning to solve scientific problems. Using forward physics informed neural networks to solve partial differential equations. Applying Backward physics informed neural networks to infer models (and PDEs) parameters.