#Approximation of Real Time Recurrent Learning