SCURank: Ranking Multiple Candidate Summaries with Summary Content Units for Enhanced Summarization

TL;DR

SCURank is a novel framework that improves summarization quality by evaluating summaries based on semantic content units, outperforming traditional metrics and LLM-based ranking methods.

Contribution

Introduces SCURank, an information-centric ranking method leveraging Summary Content Units to enhance multi-LLM distillation for summarization.

Findings

SCURank outperforms traditional metrics like ROUGE in summary ranking.

Incorporating diverse LLM summaries improves model abstractiveness.

SCURank enhances overall distilled summarization performance.

Abstract

Small language models (SLMs), such as BART, can achieve summarization performance comparable to large language models (LLMs) via distillation. However, existing LLM-based ranking strategies for summary candidates suffer from instability, while classical metrics (e.g., ROUGE) are insufficient to rank high-quality summaries. To address these issues, we introduce \textbf{SCURank}, a framework that enhances summarization by leveraging \textbf{Summary Content Units (SCUs)}. Instead of relying on unstable comparisons or surface-level overlap, SCURank evaluates summaries based on the richness and semantic importance of information content. We investigate the effectiveness of SCURank in distilling summaries from multiple diverse LLMs. Experimental results demonstrate that SCURank outperforms traditional metrics and LLM-based ranking methods across evaluation measures and datasets. Furthermore, our findings show that incorporating diverse LLM summaries enhances model abstractiveness and overall distilled model performance, validating the benefits of information-centric ranking in multi-LLM distillation. The code for SCURank is available at https://github.com/IKMLab/SCURank.