RISeC (Recipe Instruction Semantics Corpus)

该数据集包含与CURD和SIMMR相同的260烹饪食谱文本。我们的食谱指令语义语料库 (RISeC) 数据集包括零回指的文本描述，以促进其语言生成。除了数据集本身之外，我们还提供了一个流水线神经体系结构，该体系结构很好地解决了实体和关系提取的零回指。这些基本构建块可以促进更高级的下游应用程序 (例如，问题回答，会话代理)。

This dataset contains 260 cooking recipe texts identical to those used in CURD and SIMMR. Our Recipe Instruction Semantic Corpus (RISeC) dataset includes text descriptions with zero anaphora to facilitate language generation. In addition to the dataset itself, we also provide a pipeline neural architecture that effectively addresses zero anaphora for entity and relation extraction. These fundamental building blocks can enable more advanced downstream applications such as question answering and conversational agents.

搜集汇总

背景与挑战

背景概述

RISeC是一个包含260个烹饪食谱文本的数据集，专注于解决零回指问题以促进语言生成。它附带一个神经架构，用于实体和关系提取，支持问答和对话代理等下游应用。