Distributional semantics is a usage-based model of meaning, based on the assumption that the statistical distribution of linguistic items in context plays a key role in characterizing their semantic behavior. Distributional models build semantic representations by extracting co-occurrences from corpora and have become a mainstream research paradigm in computational linguistics. In this review, I present the state of the art in distributional semantics, focusing on its assets and limits as a model of meaning and as a method for semantic analysis.

中文翻译：

---

词义分布模型

分布语义是基于使用情况的意义模型，基于以下假设：上下文中语言项目的统计分布在表征其语义行为方面起着关键作用。分布模型通过从语料库中提取同现来构建语义表示，并已成为计算语言学的主流研究范式。在这篇综述中，我将介绍分布式语义的最新技术，重点是它作为意义模型和语义分析方法的资产和限制。