Abstract: A data-driven methodology is introduced for computationally efficient integration of systems of stiff rate equations in chemical kinetics using neural ordinary differential equations (NODE). A systematic algorithm is ...