CorDial: Coarse-to-fine Abstractive Dialogue Summarization with Controllable Granularity

Dialogue summarization is challenging due to its multi-speaker standpoints, casual spoken language, and limited labeled data. In this paper, we propose CorDial, aiming to improve the abstractive dialogue summarization quality and at the same time enable granularity controllability. We propose 1) a coarse-to-fine generation strategy that generates a summary draft followed by a final summary in an autoregressive way. The summary draft, which provides weakly-supervised signals, is composed of pseudo-labeled interrogative pronoun categories and noisy key phrases extracted with a constituency parser. 2) A simple strategy to control the granularity of the final summary. CorDial can predict and control the number of summary sentences for a given dialogue by predicting and highlighting different text spans from the source text. Our model achieves state-of-the-art performance on the largest dialogue summarization corpus SAMSum. We conduct comprehensive error analysis and show competitive human evaluation results to annotated summaries.