1 - If we have repeated User, Question pair with diff labels: does that mean that the user doesn't want to answer the question again and/or is it just noise ?
2 - 2 neural networks vs 1: if we have 2, we'll create a list of users first then screen that list to get the highest probabilities (is there enough data for that)
1 - If we have repeated User, Question pair with diff labels: does that mean that the user doesn't want to answer the question again and/or is it just noise ? 2 - 2 neural networks vs 1: if we have 2, we'll create a list of users first then screen that list to get the highest probabilities (is there enough data for that)