automatic keyword extraction from text

I am doing a project about automatic keyphrase extraction using sequence to sequence model by extracting web-available scientific publications as the training datasets since they contain the keywords or keyphrases assigned by the authors themselves.