None A graph-based analysis of semantic types and coercion in contextualized word embeddings

新的图方法分析词嵌入中的语义类型

作者 PulseAugur 编辑部 · [2 sources] · 2026-05-22 14:55

研究人员开发了一种新颖的基于图的方法来分析上下文词嵌入中语义类型信息的表示方式。该方法使用邻居类型概率（NTP）和邻居类型熵（NTE）等指标来检查嵌入邻域中语义类型的分布。研究发现，增强语义的嵌入能更好地捕捉词汇和上下文类型信息，从而区分具有匹配和不匹配语义类型的句子。 AI

影响引入了一个新的分析框架，用于理解词嵌入的细微差别，可能改进下游自然语言处理任务。

排序理由该集群包含一篇详细介绍新研究方法的学术论文。

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CL TIER_1 · Long Chen, Deniz Ekin Yavas · 2026-05-25 04:00

A graph-based analysis of semantic types and coercion in contextualized word embeddings

arXiv:2605.23710v1 Announce Type: new Abstract: Semantic type mismatch between a noun and its context is central to coercion phenomena. This paper introduces a graph-based method to examine how lexical and contextual type information is reflected in word embeddings. We select nou…
arXiv cs.CL TIER_1 · Deniz Ekin Yavas · 2026-05-22 14:55

A graph-based analysis of semantic types and coercion in contextualized word embeddings

Semantic type mismatch between a noun and its context is central to coercion phenomena. This paper introduces a graph-based method to examine how lexical and contextual type information is reflected in word embeddings. We select nouns from ten semantic types, annotate corpus inst…