M-Esitmator :

• Non Linear Convex Problem, to solve this problem.
• solve problem in iteraritvely optmizing by object function(cost function).
• to update parameter to do iteration, we need to have some speicifc weight to update parameter by object function.
• to avoid local minimum problem, there is many kind of ways to optimization is we called M-Estimator.
• https://en.wikipedia.org/wiki/M-estimator