NeuS: Neutral Multi-News Summarization for Mitigating Framing Bias

TL;DR

This paper introduces NeuS, a novel approach for generating neutral summaries from multiple news articles with varying political biases, aiming to reduce framing bias and promote balanced news consumption.

Contribution

The paper presents a new dataset, a novel metric, and the NeuS-TITLE model that leverages hierarchical multi-task learning to neutralize framing bias in news summarization.

Findings

NeuS-TITLE effectively reduces framing bias in generated summaries.

Hierarchical learning from titles to articles improves neutrality.

Neural models can hallucinate biased content, highlighting remaining challenges.

Abstract

Media news framing bias can increase political polarization and undermine civil society. The need for automatic mitigation methods is therefore growing. We propose a new task, a neutral summary generation from multiple news articles of the varying political leanings to facilitate balanced and unbiased news reading. In this paper, we first collect a new dataset, illustrate insights about framing bias through a case study, and propose a new effective metric and model (NeuS-TITLE) for the task. Based on our discovery that title provides a good signal for framing bias, we present NeuS-TITLE that learns to neutralize news content in hierarchical order from title to article. Our hierarchical multi-task learning is achieved by formatting our hierarchical data pair (title, article) sequentially with identifier-tokens ("TITLE=>", "ARTICLE=>") and fine-tuning the auto-regressive decoder with the standard negative log-likelihood objective. We then analyze and point out the remaining challenges and future directions. One of the most interesting observations is that neural NLG models can hallucinate not only factually inaccurate or unverifiable content but also politically biased content.