The focus of function approximation problems has been on identifying some suitable function without attempting to gain insight into the mechanism of the system. The performance of the model boils down to interpolation. But, in a more realistic setting, we expect test data from outside the distribution of the training set. To better extrapolate to unseen domains, it is essential to learn the correct underlying equations of the system. The Equation Learner (EQL) Network attempts to achieve this task.
vasisthasinghal / learning-equations-of-a-dynamical-system Goto Github PK
View Code? Open in Web Editor NEWThe focus of function approximation problems has been on identifying some suitable function without attempting to gain insight into the mechanism of the system. The performance of the model boils down to interpolation. But, in a more realistic setting, we expect test data from outside the distribution of the training set. To better extrapolate to unseen domains, it is essential to learn the correct underlying equations of the system. The Equation Learner (EQL) Network attempts to achieve this task.