GSum: A General Framework for Guided Neural Abstractive Summarization
This work addresses the issue of improving faithfulness and controllability in summarization models for NLP researchers and practitioners, representing an incremental advancement by integrating existing guidance strategies into a unified framework.
The paper tackles the problem of unfaithful and uncontrollable neural abstractive summarization by proposing GSum, a general framework that incorporates external guidance, achieving state-of-the-art ROUGE scores on four datasets and generating more faithful and controllable summaries.
Neural abstractive summarization models are flexible and can produce coherent summaries, but they are sometimes unfaithful and can be difficult to control. While previous studies attempt to provide different types of guidance to control the output and increase faithfulness, it is not clear how these strategies compare and contrast to each other. In this paper, we propose a general and extensible guided summarization framework (GSum) that can effectively take different kinds of external guidance as input, and we perform experiments across several different varieties. Experiments demonstrate that this model is effective, achieving state-of-the-art performance according to ROUGE on 4 popular summarization datasets when using highlighted sentences as guidance. In addition, we show that our guided model can generate more faithful summaries and demonstrate how different types of guidance generate qualitatively different summaries, lending a degree of controllability to the learned models.