Enhancing Long Document Long Form Summarisation with Self-Planning
This work addresses the challenge of maintaining accuracy and detail in long-form summarization for applications like document analysis, though it appears incremental as it builds on self-planning methods.
The paper tackles the problem of generating faithful and traceable summaries for long documents by introducing a highlight-guided generation approach that uses sentence-level content planning. Their method improves factual consistency, achieving a 4.1-point gain in ROUGE-L and about 35% gains in SummaC scores on the GovReport dataset.
We introduce a novel approach for long context summarisation, highlight-guided generation, that leverages sentence-level information as a content plan to improve the traceability and faithfulness of generated summaries. Our framework applies self-planning methods to identify important content and then generates a summary conditioned on the plan. We explore both an end-to-end and two-stage variants of the approach, finding that the two-stage pipeline performs better on long and information-dense documents. Experiments on long-form summarisation datasets demonstrate that our method consistently improves factual consistency while preserving relevance and overall quality. On GovReport, our best approach has improved ROUGE-L by 4.1 points and achieves about 35% gains in SummaC scores. Qualitative analysis shows that highlight-guided summarisation helps preserve important details, leading to more accurate and insightful summaries across domains.