Invited Talk: On Proactive Sentence Specific Popularity Forecasting

Abstract

In this draft, we introduce the reader to the problem of popularity prediction. We highlight the existing directions of work in the area, and lay out the foundation for the task of proactively forecasting relative information popularity of individual sentences within online news documents. We discuss the key challenges and potential business applications for this novel task and note down the main contributions of our work presented at HT ’22 [7]. Lastly, we discuss some interesting avenues of future work.