Towards Proactively Forecasting Sentence-Specific Information Popularity within Online News Documents

TL;DR

This paper introduces a new task of predicting the popularity of individual sentences in online news articles using natural language content, supported by a novel dataset and transfer learning approach.

Contribution

It presents the first dataset for sentence-level popularity prediction and a transfer learning method leveraging salience prediction to improve forecasting accuracy.

Findings

Achieved nDCG > 0.8 in popularity forecasting

Transfer learning from salience prediction improves performance

First dataset with sentence-level popularity labels from search queries

Abstract

Multiple studies have focused on predicting the prospective popularity of an online document as a whole, without paying attention to the contributions of its individual parts. We introduce the task of proactively forecasting popularities of sentences within online news documents solely utilizing their natural language content. We model sentence-specific popularity forecasting as a sequence regression task. For training our models, we curate InfoPop, the first dataset containing popularity labels for over 1.7 million sentences from over 50,000 online news documents. To the best of our knowledge, this is the first dataset automatically created using streams of incoming search engine queries to generate sentence-level popularity annotations. We propose a novel transfer learning approach involving sentence salience prediction as an auxiliary task. Our proposed technique coupled with a BERT-based neural model exceeds nDCG values of 0.8 for proactive sentence-specific popularity forecasting. Notably, our study presents a non-trivial takeaway: though popularity and salience are different concepts, transfer learning from salience prediction enhances popularity forecasting. We release InfoPop and make our code publicly available: https://github.com/sayarghoshroy/InfoPopularity