发表机构
University of Toronto; University of Waterloo（多伦多大学; 滑铁卢大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出无需训练的LOOMSUM框架，结合Table-Grounded Faithfulness指标，在FINDSum和USTT基准上提升长文本-表格摘要的分析忠实性，显式跨模态关联可减少事实配对错误。
AI 中文摘要
长文档常将重要信息分散在大量叙事段落和多个表格中，这使得忠实摘要的生成极具挑战性。现有方法可能生成各自有依据的定量事实与分析性陈述，却将它们错误关联，产生定量看似合理但分析层面不忠实的摘要。本研究提出LOOMSUM，这是一个无需训练的框架，它提取基于源文档的原子证据，明确关联表格衍生事实与支持性叙事分析，并在生成前规划话语结构。我们还引入了Table-Grounded Faithfulness（TGF），这是一种主张级指标，分别评估数值依据、分析支持和关系一致性。在文本-表格摘要基准FINDSum和USTT上的实验表明，LOOMSUM在保持较强摘要质量的同时提升了分析忠实性。人工评估发现其组件级与对应人工判断存在正相关。我们的关系一致性指标与通用事实性指标相比，与人工关系判断的一致性更强，这表明显式跨模态关联有助于减少支持的数量与错误叙事解读配对的错误。综上，这些发现表明，忠实的长文本-表格摘要不仅需要锚定单个事实，还需保留事实间的关系。
英文摘要
Long documents often distribute important information across extensive narrative passages and multiple tables, making faithful summarization particularly challenging. Existing methods may generate individually supported quantitative facts and analytical statements yet associate them incorrectly, producing quantitatively plausible yet analytically unfaithful summaries. In this work, we propose LOOMSUM, a training-free framework that extracts source-grounded atomic evidence, explicitly links table-derived facts with supporting narrative analyses, and plans the discourse structure before generation. We also introduce Table-Grounded Faithfulness (TGF), a claim-level metric that separately evaluates Numeric Grounding, Analysis Support, and Relation Consistency. Experiments on the text--table summarization benchmarks FINDSum and USTT show that LOOMSUM improves analytical faithfulness while maintaining strong summarization quality. Human evaluation finds positive component-level associations with the corresponding human judgments. Our Relation Consistency metric further shows stronger agreement with human relation judgments than generic factuality metrics, indicating that explicit cross-modal linking helps reduce errors in which supported quantities are paired with incorrect narrative interpretations. Together, these findings show that faithful long text--table summarization requires not only grounding individual facts, but also preserving the relations between them.
CommentsPreprint, code will be available soon
