Text this: Estimation of the Conditional Probability Using a Stochastic Gradient Process