Argument Generation with Retrieval, Planning, and Realization

TL;DR

This paper introduces CANDELA, a novel framework for counter-argument generation that combines retrieval from a large article database with a two-step generation process involving planning and realization, resulting in more relevant and content-rich arguments.

Contribution

The paper presents a new retrieval-augmented argument generation model with a two-step decoding process, improving the quality and diversity of generated counter-arguments.

Findings

Outperforms state-of-the-art models on BLEU, ROUGE, and METEOR scores.

Generates more appropriate and content-rich counter-arguments according to human evaluation.

Uses a retrieval system with 12 million articles for high-quality content access.

Abstract

Automatic argument generation is an appealing but challenging task. In this paper, we study the specific problem of counter-argument generation, and present a novel framework, CANDELA. It consists of a powerful retrieval system and a novel two-step generation model, where a text planning decoder first decides on the main talking points and a proper language style for each sentence, then a content realization decoder reflects the decisions and constructs an informative paragraph-level argument. Furthermore, our generation model is empowered by a retrieval system indexed with 12 million articles collected from Wikipedia and popular English news media, which provides access to high-quality content with diversity. Automatic evaluation on a large-scale dataset collected from Reddit shows that our model yields significantly higher BLEU, ROUGE, and METEOR scores than the state-of-the-art and non-trivial comparisons. Human evaluation further indicates that our system arguments are more appropriate for refutation and richer in content.