How to classify text data with categorical data using deep learning (cnn, lstm, etc...)

Im trying to classify text data with categorical data. (with deep learning)

The issue is data with a category must return in specific targets.

example)

category 1 labels in (a,b,c)

category 2 labels in (a,b,d)

category 3 labels in (b,c,d,e)

also, the texts are very similar across all categories

Is there any way to train with this data?