A Statistical Analysis of Summarization Evaluation Metrics using Resampling Methods
Preprint
Abstract
The quality of a summarization evaluation metric is quantified by calculating the correlation between its scores and human annotations across a large number of summaries. Currently, it is unclear how precise these correlation estimates are, nor whether differences between two metrics' correlations reflect a true difference or if it is due to mere chance. In this work, we address these two problems by proposing methods for calculating confidence intervals and running hypothesis tests for correlations using two resampling methods, bootstrapping and permutation. After evaluating which of the proposed methods is most appropriate for summarization through two simulation experiments, we analyze the results of applying these methods to several different automatic evaluation metrics across three sets of human annotations. We find that the confidence intervals are rather wide, demonstrating high uncertainty in the reliability of automatic metrics. Further, although many metrics fail to show statistical improvements over ROUGE, two recent works, QAEval and BERTScore, do in some evaluation settings.Keywords
All Related Versions
This publication has 17 references indexed in Scilit:
- Towards Question-Answering as an Automatic Metric for Evaluating the Content Quality of a SummaryTransactions of the Association for Computational Linguistics, 2021
- SummEval: Re-evaluating Summarization EvaluationTransactions of the Association for Computational Linguistics, 2021
- SacreROUGE: An Open-Source Library for Using and Developing Summarization Evaluation Metrics2020
- Re-evaluating Evaluation in Text Summarization2020
- Not All Claims are Created Equal: Choosing the Right Statistical Approach to Assess Hypotheses2020
- Replicability Analysis for Natural Language Processing: Testing Significance with Multiple DatasetsTransactions of the Association for Computational Linguistics, 2017
- Summary Evaluation: Together We Stand NPowER-ed2013
- Summarization system evaluation revisitedACM Transactions on Speech and Language Processing, 2008
- Sample Size Requirements for Estimating Pearson, Kendall and Spearman CorrelationsPsychometrika, 2000
- An Introduction to the Bootstrap1993