On The Stability Of Inverse Problems