I wonder how can I build a neural network which will generate text description from given tag/tags. Let's assume I have created such data structure:

 'tag1': ['some description1', 'some description2', 'some description3'],
 'tag2': ['some description4', 'some description5', 'some description6'],
 'tag3': ['some description7', 'some description8', 'some description9']

Then I would like to create a neural network which will generate randomly generated description based on given tags. For example:

INPUT: ['TAG1', 'TAG2', 'TAG3'] => OUTPUT: 'some description1. some description5 some description9'

Then I thought that it can be a good idea to implement a LSTM and doing text generation, but here I have a problem I know how I can do it for one tag. I can create one corpus of text contains different sentences for tag, then do the training and generate a sentence for given tag, but what If I have multiple tags should I create a corpus for each tag or maybe there is a better way to do that? If you know any articles which covers this problem, I would appreciate if you share them with me. If you have a neural network proposition which will solve this problem, I am also open for proposals.

PS. I know, I can solve this problem with easy Map, for example: ['tag1', 'tag2', 'tag3'].map(tag => tagSentenceMap.get(tag).randomChoice()).join('. ') but this is not the case for me.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.