My main goal is to come up with some topics using LSTM autoencoder.
I want to use 20 news_group data set.
after reading lots of material and looking at some GitHub project, I am still not clear how to prepare my data for the goal I am following. particularly,
what will be the label here? (If I want to feed each paragraph, should the label will be the next paragraph?)
Most of the project is having an embedding in the first layer, then what will be the difference if feeding the model with a sequence of text, like two sentences as a sequence rather having an embedding in the model? (I think I am confused with the concept of having embedding layer VS fitting the input as a sequence of ordered words for two sentences in the first layer.)