Persian Social Norms (PSN)

Persian Social Norms (PSN)数据集是由米兰理工大学和米兰大学共同创建，旨在捕捉伊朗社会的文化规范，以支持全球意识AI系统的开发。该数据集包含1.7K条伊朗社会规范，涵盖环境、情境和文化标签，并附有英文翻译。数据集的创建过程利用了大型语言模型和提示工程技术，生成的规范由母语为波斯语的专家进行质量与伦理合规性审查。PSN数据集作为首个此类波斯语数据集，支持跨文化AI中的规范适应计算建模，是一个关键的跨文化挑战，由多元文化视角提供信息。数据集的应用领域包括跨文化研究和支持开发更具包容性的AI系统，旨在解决AI系统在不同文化背景下的适应性和理解问题。

The Persian Social Norms (PSN) dataset was co-developed by Politecnico di Milano and the University of Milan, aiming to capture the cultural norms of Iranian society to support the development of globally aware AI systems. This dataset contains 1.7 thousand Iranian social norm entries, covering environmental, situational and cultural tags, with accompanying English translations. The development of the dataset utilized large language models (LLMs) and prompt engineering technologies, and the generated social norms were reviewed for quality and ethical compliance by native Persian-speaking experts. As the first Persian-language dataset of its kind, the PSN dataset enables computational modeling of norm adaptation in cross-cultural AI—a critical cross-cultural challenge informed by multicultural perspectives. Its application fields include cross-cultural research and supporting the development of more inclusive AI systems, which aim to address the adaptability and understanding issues of AI systems across different cultural backgrounds.