linagora/MinecraftStructuredDialogueCorpus

Minecraft结构化对话语料库（MSDC）是基于Minecraft对话语料库（MDC）的一个话语标注版本。该数据集包含了人类参与者在模拟的3D Minecraft网格环境中，通过聊天窗口进行的自然语言对话。对话中，一方为建筑师，另一方为建造者。建筑师通过键入指令来指导建造者在一个网格上重建一个结构。该语料库由话语专家和学生小组进行了标注，将对话划分为基本话语单位（EDUs）和建造者的事件单位（EEUs），并记录了话语关系实例。数据集分为训练集、验证集、测试集和开发集，共有540个对话。

The Minecraft Structured Dialogue Corpus (MSDC) is a discourse-annotated version of the Minecraft Dialogue Corpus (MDC). It contains natural language dialogues between human participants in a simulated 3D Minecraft grid environment, communicating through a chat window. One participant acts as the Architect, and the other as the Builder. The Architect aims to guide the Builder to recreate a structure on the grid through typed instructions. The corpus has been annotated by a group of discourse experts and students, segmenting the dialogues into elementary discourse units (EDUs) and Builder action units (EEUs), and recording discourse relation instances. The dataset is divided into training, validation, test, and development sets, totaling 540 dialogues.