Extending the machine-learned optimized stable Taylor-Galerkin scheme to solve the Burgers' equation