What is the advantage of Variable step size LMS over Leaky-LMS adaptive filter algorithm? Which one has a better performance?
Variable step size LMS is generally used to improve the speed of convergence or decrease steady-state error. Leaky adaptation is used to combat problems like the potential instability of the filter in a finite-precision implementation. It is closely related to the L2 norm regularization technique and results in continuous downscaling of filter coefficients (hence smoothing the extracted filter). So comparing these two techniques is not that meaningful. In fact, you can easily use both techniques together to achieve all the above advantages together.