SherLIiC

SherLIiC 是上下文词汇推理 (LIiC) 的测试平台，由 3985 个手动注释的推理规则候选 (InfCands) 组成，伴随着 (i) ~960k 未标记的 InfCands，以及 (ii) 从大型实体链接语料库 ClueWeb09.每个 InfCand 由这些关系之一组成，表示为一个词形还原的依赖路径，以及两个参数占位符，每个占位符都链接到一个或多个 Freebase 类型。SherLIiC is a testbed for contextual lexical inference (LIiC), comprising 3985 manually annotated inference rule candidates (InfCands), accompanied by (i) approximately 960k unlabeled InfCands, and (ii) candidates extracted from the large entity-linking corpus ClueWeb09. Each InfCand consists of one of these relations, represented as a lemmatized dependency path, along with two argument placeholders, each linked to one or more Freebase types.

搜集汇总

背景与挑战

背景概述

SherLIiC是一个用于评估自然语言推理的基准数据集，包含3985个手动标注的推理规则候选和约96万个未标记候选，数据源自ClueWeb09实体链接语料库。每个候选由词形还原的依赖路径和链接到Freebase类型的参数占位符组成，旨在支持上下文词汇推理任务。