NLP-CAutoCabCC: A Chinese Automobile Cabin Command Corpus

该开源数据集包含500句中文座舱功能点泛化语料，涵盖10种车载命令控制功能，每个功能相关含10-100种通用语料。 如开启车道保持、开启遮阳帘、打开远光灯、打开蓝牙、打开WiFi、启动ESP等等。句式多样性丰富，在语句结构上充分考虑了动词、实体词、句式及其组合，同时对功能点的多样性表达进行泛化，如车身稳定系统=ESP，延时摄像=缩时录像。在泛化部件功能时文本有预留Slot，涉及槽位有Position、Fraction、Percent等，如position=[前，后，左，右，中，左后方，全部]等。This open-source dataset contains 500 Chinese in-vehicle cockpit function point generalization utterances, covering 10 types of in-vehicle command and control functions, with 10 to 100 general utterances per function. Examples include: activate lane keeping, open sunshade, turn on high beam, enable Bluetooth, turn on WiFi, activate ESP, etc. The dataset features rich sentence diversity, with full consideration given to verbs, entity terms, sentence structures and their combinations during utterance formulation. Meanwhile, it generalizes diverse expressions of function points; for example, "Vehicle Stability Control System" equals "ESP", and "time-lapse photography" equals "time-lapse recording". When generalizing component functions, the text reserves pre-defined Slots, with involved Slot types including Position, Fraction, Percent, etc. For instance, the Position Slot has valid values such as [front, rear, left, right, middle, rear-left, all].

搜集汇总

背景与挑战

背景概述

该数据集是一个中文汽车座舱命令语料库，包含500句泛化语料，覆盖10种车载控制功能，如开启车道保持、打开蓝牙等。其特点是句式多样性丰富，对功能点进行泛化表达，并预留了Slot槽位以支持部件功能泛化，适用于自然语言理解和自动驾驶相关研究。