Reducing model bias in a deep learning classifier using domain adversarial neural networks in the MINERvA experiment. (26th November 2018)