Commonsense Explanation Arabic Dataset

Commonsense Explanation Arabic Dataset是由约旦科技大学计算机科学系创建的一个基准阿拉伯语数据集，旨在评估机器对阿拉伯语句子的常识解释能力。该数据集包含12000条记录，分为训练集、验证集和测试集，每条记录包含一个不合逻辑的句子及其三个可能的解释，任务是选择最合理的解释。数据集的创建过程涉及从英语数据集翻译和改编，以适应阿拉伯语环境。该数据集主要用于自然语言处理领域，特别是常识推理和语言理解的研究，旨在提高机器对阿拉伯语常识问题的解析能力。

Commonsense Explanation Arabic Dataset is a benchmark Arabic dataset developed by the Department of Computer Science at Jordan University of Science and Technology, designed to evaluate machines' commonsense explanation capabilities for Arabic sentences. This dataset contains 12,000 records divided into training, validation and test subsets. Each record includes an illogical sentence and three candidate explanations, with the task of selecting the most plausible one. The dataset was constructed by translating and adapting an existing English dataset to suit the Arabic linguistic and contextual environment. It is primarily applied in natural language processing (NLP) research, particularly for studies on commonsense reasoning and language understanding, with the goal of enhancing machines' ability to process and comprehend Arabic commonsense-related problems.