The Computational Physics and Machine Learning lab fosters the use of machine learning and advanced computational techniques for the data-driven modeling of complex systems.  Special focus is on the development of novel algorithms that can discover the governing equations underlying some phenomena starting from experimental measurements or other type of data. More broadly, we strive to advance Scientific Machine Learning with a combination of symbolic and neural approaches.

In contrast with purely data-driven methods and current trends in AI and ML, we strive to obtain trustworthy models provide guarantees on their safety and reliability, and generalize well on unseen data. To achieve these goals, we use a combination of equation-based modeling and machine learning techniques, including symbolic regression, evolutionary algorithms, reinforcement learning and deep neural networks.