Hello! I’m a newbie in the NLP, so currently I’m struggling with the following problem. Let’s assume, that on the input I have text like this: "The price of apple juice has increased" and text in output like this: "Increase the production of apple juice". Using Recurrent Neural Network, can the output be predicted? What kind of algorithm I should use for it?
Seems like Google’s BERT will be the best choice for you. BERT is pre-trained on "Next Sentence Prediction" and allows to have two separate sentences as input which can help a lot for solving your task.
Look at the seq2seq models. In general, you can use one RNN as an encoder and another RNN as a decoder that will take the encoded vector as in input to generate a new sequence of words.
Thank you.
Just drop us an email to ...
Show more 
Log in with Google