Direct-Scoring NLG Evaluators Can Use Pairwise Comparisons Too

• URL: http://arxiv.org/abs/2509.05440v1
• Date: Fri, 05 Sep 2025 18:48:34 GMT
• Title: Direct-Scoring NLG Evaluators Can Use Pairwise Comparisons Too
• Authors: Logan Lawrence, Ashton Williamson, Alexander Shelton,
• Abstract summary: We propose a direct-scoring method which uses synthetic summaries to act as pairwise machine rankings at test time.<n>We show that our method performs comparably to state-of-the-art pairwise evaluators in terms of axis-averaged sample-level correlations.
• Score: 42.13843953705695
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As large-language models have been increasingly used as automatic raters for evaluating free-form content, including document summarization, dialog, and story generation, work has been dedicated to evaluating such models by measuring their correlations with human judgment. For \textit{sample-level} performance, methods which operate by using pairwise comparisons between machine-generated text perform well but often lack the ability to assign absolute scores to individual summaries, an ability crucial for use cases that require thresholding. In this work, we propose a direct-scoring method which uses synthetic summaries to act as pairwise machine rankings at test time. We show that our method performs comparably to state-of-the-art pairwise evaluators in terms of axis-averaged sample-level correlations on the SummEval (\textbf{+0.03}), TopicalChat (\textbf{-0.03}), and HANNA (\textbf{+0.05}) meta-evaluation benchmarks, and release the synthetic in-context summaries as data to facilitate future work.

Related papers

• Pairwise Comparison for Bias Identification and Quantification [5.490018587818889]
  Linguistic bias in online news and social media is difficult to measure.<n>We aim to reduce annotation effort by leveraging pairwise comparison for bias annotation.
  arXiv  Detail & Related papers  (2025-12-16T16:36:55Z)
• A Comparative Study of Quality Evaluation Methods for Text Summarization [0.5512295869673147]
  This paper proposes a novel method based on large language models (LLMs) for evaluating text summarization. Our results show that LLMs evaluation aligns closely with human evaluation, while widely-used automatic metrics such as ROUGE-2, BERTScore, and SummaC do not and also lack consistency.
  arXiv  Detail & Related papers  (2024-06-30T16:12:37Z)
• Long-Span Question-Answering: Automatic Question Generation and QA-System Ranking via Side-by-Side Evaluation [65.16137964758612]
  We explore the use of long-context capabilities in large language models to create synthetic reading comprehension data from entire books. Our objective is to test the capabilities of LLMs to analyze, understand, and reason over problems that require a detailed comprehension of long spans of text.
  arXiv  Detail & Related papers  (2024-05-31T20:15:10Z)
• Is Summary Useful or Not? An Extrinsic Human Evaluation of Text Summaries on Downstream Tasks [45.550554287918885]
  This paper focuses on evaluating the usefulness of text summaries with extrinsic methods. We design three different downstream tasks for extrinsic human evaluation of summaries, i.e., question answering, text classification and text similarity assessment. We find summaries are particularly useful in tasks that rely on an overall judgment of the text, while being less effective for question answering tasks.
  arXiv  Detail & Related papers  (2023-05-24T11:34:39Z)
• Evaluating Factual Consistency of Texts with Semantic Role Labeling [3.1776833268555134]
  We introduce SRLScore, a reference-free evaluation metric designed with text summarization in mind. A final factuality score is computed by an adjustable scoring mechanism. Correlation with human judgments on English summarization datasets shows that SRLScore is competitive with state-of-the-art methods.
  arXiv  Detail & Related papers  (2023-05-22T17:59:42Z)
• SWING: Balancing Coverage and Faithfulness for Dialogue Summarization [67.76393867114923]
  We propose to utilize natural language inference (NLI) models to improve coverage while avoiding factual inconsistencies. We use NLI to compute fine-grained training signals to encourage the model to generate content in the reference summaries that have not been covered. Experiments on the DialogSum and SAMSum datasets confirm the effectiveness of the proposed approach.
  arXiv  Detail & Related papers  (2023-01-25T09:33:11Z)
• On the Limitations of Reference-Free Evaluations of Generated Text [64.81682222169113]
  We show that reference-free metrics are inherently biased and limited in their ability to evaluate generated text. We argue that they should not be used to measure progress on tasks like machine translation or summarization.
  arXiv  Detail & Related papers  (2022-10-22T22:12:06Z)
• Comparing Methods for Extractive Summarization of Call Centre Dialogue [77.34726150561087]
  We experimentally compare several such methods by using them to produce summaries of calls, and evaluating these summaries objectively. We found that TopicSum and Lead-N outperform the other summarisation methods, whilst BERTSum received comparatively lower scores in both subjective and objective evaluations.
  arXiv  Detail & Related papers  (2022-09-06T13:16:02Z)
• SMART: Sentences as Basic Units for Text Evaluation [48.5999587529085]
  In this paper, we introduce a new metric called SMART to mitigate such limitations. We treat sentences as basic units of matching instead of tokens, and use a sentence matching function to soft-match candidate and reference sentences. Our results show that system-level correlations of our proposed metric with a model-based matching function outperforms all competing metrics.
  arXiv  Detail & Related papers  (2022-08-01T17:58:05Z)
• SummEval: Re-evaluating Summarization Evaluation [169.622515287256]
  We re-evaluate 14 automatic evaluation metrics in a comprehensive and consistent fashion. We benchmark 23 recent summarization models using the aforementioned automatic evaluation metrics. We assemble the largest collection of summaries generated by models trained on the CNN/DailyMail news dataset.
  arXiv  Detail & Related papers  (2020-07-24T16:25:19Z)
• Evaluating Text Coherence at Sentence and Paragraph Levels [17.99797111176988]
  We investigate the adaptation of existing sentence ordering methods to a paragraph ordering task. We also compare the learnability and robustness of existing models by artificially creating mini datasets and noisy datasets. We conclude that the recurrent graph neural network-based model is an optimal choice for coherence modeling.
  arXiv  Detail & Related papers  (2020-06-05T03:31:49Z)
• Learning by Semantic Similarity Makes Abstractive Summarization Better [13.324006587838522]
  We compare the generated summaries from recent LM, BART, and the reference summaries from a benchmark dataset, CNN/DM. Interestingly, model-generated summaries receive higher scores relative to reference summaries.
  arXiv  Detail & Related papers  (2020-02-18T17:59:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.