A Methodology Establishing Linear Convergence of Adaptive Gradient Methods under PL Inequality | Read Paper on Bytez