Normally when you summarize text using a model like the facebook/bart-large-cnn
or the sshleifer/distilbart-cnn-12-6
it generates a general summary based on the main ides of the text, i want to know how you would be able to target a single word or topic, and have the model summarize the text around that word, instead of the general summary. Any help would really be appreciated.
I’ve only skimmed this paper, but maybe it could be done with something like what’s described here: A Reinforced Topic-Aware Convolutional Sequence-to-Sequence Model for Abstractive Text Summarization. Or at the very least it might point you in a useful direction! Here’s a Connected Papers graph for more reading: Connected Papers | Find and explore academic papers
1 Like