# Embrace Divergence for Richer Insights: A Multi-document Summarization Benchmark and a Case Study on Summarizing Diverse Information from News Articles Source: [Embrace Divergence for Richer Insights: A Multi-document Summarization Benchmark and a Case Study on Summarizing Diverse Information from News Articles]() Collection index: [llms.txt]() Data updated: 2025-10-26T02:30:51.475Z Authors: Kung-Hsiang Huang, Philippe Laban, Alexander R. Fabbri, Prafulla Kumar Choubey, Shafiq Joty, Caiming Xiong, Chien-Sheng Wu Published: 2023-09-17T20:28:17.000Z ## Abstract Previous research in multi-document news summarization has typically concentrated on collating information that all sources agree upon. However, the summarization of diverse information dispersed across multiple articles about an event remains underexplored. In this paper, we propose a new task of summarizing diverse information encountered in multiple news articles encompassing the same event. To facilitate this task, we outlined a data collection schema for identifying diverse information and curated a dataset named DiverseSumm. The dataset includes 245 news stories, with each story comprising 10 news articles and paired with a human-validated reference. Next, to enable consistent automatic evaluation, we conducted a comprehensive analysis to pinpoint the position and verbosity biases when utilizing Large Language Model (LLM)-based metrics for evaluating the coverage and faithfulness of summaries. Through correlation analyses, we outline the best practices for effectively using automatic LLM-based metrics on the DiverseSumm dataset. Finally, we study how LLMs summarize multiple news articles by analyzing which type of diverse information LLMs are capable of identifying. Our analyses suggest that despite the extraordinary capabilities of LLMs in single-document summarization, the proposed task remains a complex challenge for them mainly due to their limited coverage, with GPT-4 only able to cover under 40% of the diverse information on average. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Source Coverage Evaluation on DIVERSESUMM 1.0 (test)](): Source Coverage Evaluation; DIVERSESUMM 1.0 (test); Source Coverage: 55.16. - [Source Coverage Evaluation on DIVERSESUMM Augmented 1.0 (test)](): Source Coverage Evaluation; DIVERSESUMM Augmented 1.0 (test); Source Coverage: 0.4807. - [Query-focused Multi-document Summarization on DIVERSESUMM Relevance](): Query-focused Multi-document Summarization; DIVERSESUMM Relevance; Coverage: 64.1. ## Original sources and implementation [Original paper]() [PDF]()