Dialogue Inspectional Summarization with Factual Inconsistency Awareness

TL;DR

This paper addresses factual inconsistencies in dialogue summarization, proposing an end-to-end framework with auxiliary tasks to improve factual accuracy and detect missing facts, especially in professional dialogues.

Contribution

It introduces a novel framework with EFAR and MFED tasks to enhance factual consistency in dialogue summarization without relying on pretraining.

Findings

Improved factual accuracy in generated summaries.

Effective detection of missing factual entities.

Enhanced readability and factual coverage.

Abstract

Dialogue summarization has been extensively studied and applied, where the prior works mainly focused on exploring superior model structures to align the input dialogue and the output summary. However, for professional dialogues (e.g., legal debate and medical diagnosis), semantic/statistical alignment can hardly fill the logical/factual gap between input dialogue discourse and summary output with external knowledge. In this paper, we mainly investigate the factual inconsistency problem for Dialogue Inspectional Summarization (DIS) under non-pretraining and pretraining settings. An innovative end-to-end dialogue summary generation framework is proposed with two auxiliary tasks: Expectant Factual Aspect Regularization (EFAR) and Missing Factual Entity Discrimination (MFED). Comprehensive experiments demonstrate that the proposed model can generate a more readable summary with accurate coverage of factual aspects as well as informing the user with potential missing facts detected from the input dialogue for further human intervention.