search

shlomihod / deep-text-eval

Differnable Readability Measure Regularizer for Neural Network Automatic Text Simplification
24 stars 7 forks source link

Use 2 stacked RNN: one for learning the NLP features, the other on classification #25

Open shlomihod opened 6 years ago

shlomihod commented 6 years ago

Use the temperature idea to make the distribution very cold on the text2features NN, and then getting something close to a count (for POS)