NewsDialogues: Towards Proactive News Grounded Conversation

TL;DR

This paper introduces a new task of proactive news grounded conversation, creates a Chinese dialogue dataset, and proposes a method to improve dialogue systems' ability to lead conversations based on news topics.

Contribution

It defines a novel proactive news grounded conversation task, constructs a Chinese dataset, and develops the Predict-Generate-Rank method to enhance dialogue system performance.

Findings

The proposed method improves response relevance and coherence.

The dataset enables realistic evaluation of news-grounded dialogues.

Challenges in maintaining engaging and accurate news conversations are identified.

Abstract

Hot news is one of the most popular topics in daily conversations. However, news grounded conversation has long been stymied by the lack of well-designed task definition and scarce data. In this paper, we propose a novel task, Proactive News Grounded Conversation, in which a dialogue system can proactively lead the conversation based on some key topics of the news. In addition, both information-seeking and chit-chat scenarios are included realistically, where the user may ask a series of questions about the news details or express their opinions and be eager to chat. To further develop this novel task, we collect a human-to-human Chinese dialogue dataset \ts{NewsDialogues}, which includes 1K conversations with a total of 14.6K utterances and detailed annotations for target topics and knowledge spans. Furthermore, we propose a method named Predict-Generate-Rank, consisting of a generator for grounded knowledge prediction and response generation, and a ranker for the ranking of multiple responses to alleviate the exposure bias. We conduct comprehensive experiments to demonstrate the effectiveness of the proposed method and further present several key findings and challenges to prompt future research.