AkihikoWatanabe / paper_notes

たまに追加される論文メモ
https://AkihikoWatanabe.github.io/paper_notes
17 stars 0 forks source link

ROUGE-C: A fully automated evaluation method for multi-document summarization, He+, International Conference on Granular Computing'08 #955

Open AkihikoWatanabe opened 1 year ago

AkihikoWatanabe commented 1 year ago

https://ieeexplore.ieee.org/abstract/document/4664680

AkihikoWatanabe commented 1 year ago

This paper presents how to use ROUGE to evaluate summaries without human reference summaries. ROUGE is a widely used evaluation tool for multi-document summarization and has great advantages in the areas of summarization evaluation. However, manual reference summaries written beforehand by assessors are indispensable for a ROUGE test. There was still no research on ROUGEpsilas abilities of evaluating summaries without manual reference summaries. By considering summary as consensus speaker for the original input information, we discovered and developed ROUGE-C. ROUGE-C applies the ROUGE method alternatively by replacing the reference summaries with source document as well as query-focused information (if any), and therefore it enables a fully manual-independent way of evaluating multi-document summarization. Experiments conducted on the 2001 to 2005 DUC data showed that, with restraint of appropriate condition and some acceptable decreased efficiency, ROUGE-C correlated well with methods that depend on reference summaries, including human judgments.

Translation (by gpt-3.5-turbo)