Improving the Faithfulness of Abstractive Summarization via Entity Coverage Control
This addresses the issue of unfaithful summaries for users relying on automated systems, though it is incremental as it builds on existing pre-trained models.
The paper tackled the problem of fact hallucination in abstractive summarization by proposing Entity Coverage Control (ECC), which improved faithfulness and saliency in both supervised fine-tuning and zero-shot settings across three benchmark datasets.
Abstractive summarization systems leveraging pre-training language models have achieved superior results on benchmark datasets. However, such models have been shown to be more prone to hallucinate facts that are unfaithful to the input context. In this paper, we propose a method to remedy entity-level extrinsic hallucinations with Entity Coverage Control (ECC). We first compute entity coverage precision and prepend the corresponding control code for each training example, which implicitly guides the model to recognize faithfulness contents in the training phase. We further extend our method via intermediate fine-tuning on large but noisy data extracted from Wikipedia to unlock zero-shot summarization. We show that the proposed method leads to more faithful and salient abstractive summarization in supervised fine-tuning and zero-shot settings according to our experimental results on three benchmark datasets XSum, Pubmed, and SAMSum of very different domains and styles.