Is the Prediction a Challenge?

The model is aimed to optimize the fact that the score they predict to be the positive sample has to be larger than the score that predicted e the negative sample. One danger if we're going to apply deep learning techniques is that there could be some sort of over fitting going on. But what will make it actually useful is turning those not human readable feature victors and making them some action items.