AI Data Readiness Inspector (AIDRIN) for Quantitative Assessment of Data Readiness for AI

“垃圾进，垃圾出”是计算机科学家们普遍认同的格言，包括人工智能（AI）领域的科学家在内。由于数据是AI的燃料，基于低质量、有偏见的数据训练的模型通常是无效的。使用AI的计算机科学家在为AI准备数据方面投入了大量的时间和精力。然而，目前没有标准的方法或框架来评估数据是否“准备好”用于AI。为了提供数据准备就绪的可量化评估，我们定义了AI数据准备就绪的参数，并引入了AIDRIN（AI数据准备就绪检查器）。AIDRIN是一个框架，涵盖了文献中可用于定量和定性评估数据准备就绪的广泛维度。AIDRIN使用传统数据质量评估中的度量标准，如完整性、异常值和重复项，用于数据评估。此外，AIDRIN使用特定的度量标准来评估用于AI的数据，如特征重要性、特征相关性、类别不平衡、公平性、隐私和FAIR（可发现性、可访问性、互操作性和可重用性）原则的合规性。AIDRIN提供可视化和报告，以帮助数据科学家进一步调查数据的准备就绪情况。AIDRIN框架提高了机器学习流程的效率，以便对数据准备就绪情况做出明智的决策，应用于AI应用程序。

AIDRIN: Quantitative and Qualitative AI Data Readiness Inspection Framework

AIDRIN is a framework that provides a quantifiable assessment of the readiness of data for AI processes by defining parameters of AI data readiness and using metrics specific to assess data for AI, such as feature importance, fairness, and privacy.

AIDRIN uses metrics in traditional data quality assessment and provides visualizations and reports to assist data scientists in further investigating the readiness of data. The framework enhances the efficiency of the machine learning pipeline to make informed decisions on data readiness for AI applications.

Related studies in this field include 'Data Quality for Machine Learning: Some Definitions and Issues' by Batista et al. and 'Towards a Rigorous Science of Interpretable Machine Learning' by Doshi-Velez and Kim.