English(EN) Multi-Label Agreement: A Flexible Framework for Inter-Annotator Reliability

新框架MLA增强了多标签任务中标注者间的一致性可靠性

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-28 04:00

研究人员推出了一种名为多标签一致性（MLA）的新框架，旨在改进多标签标注任务中标注者间可靠性的评估。与主要用于单标签场景的现有指标不同，MLA考虑了分配给每个项目的完整二元标签向量，通过平均成对汉明一致性来考虑联合选择和遗漏。该框架通过汇集边际、Gwet AC1和固定基线归一化提供了灵活性，并提供项目、标签和语料库级别的详细结果，以及条件最小一致性诊断。提供了Python实现，可进行透明且可解释的多标签一致性分析。 AI

影响增强了数据标注的可靠性，这是开发和评估AI模型的关键步骤。

排序理由该项目是一篇介绍评估标注者间一致性新框架的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 15 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该项目是一篇介绍评估标注者间一致性新框架的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, other; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Breaking (< 6h); Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Tim Menzner, Jochen L. Leidner · 2026-10-28 04:00

多标签协议：一个用于标注者间可靠性的灵活框架

arXiv:2610.37292v1 Announce Type: new Abstract: Reliable evaluation of inter-annotator agreement (IAA) is important for assessing the consistency of annotation processes. While established IAA metrics such as Cohen's kappa, Fleiss' kappa, and Krippendorff's alpha are commonly app…

本文如何被排名

报道来源 [1]

多标签协议：一个用于标注者间可靠性的灵活框架

相关实体

相关话题