Concepts in Quality Assessment for Machine Learning - From Test Data to Arguments
Concepts in Quality Assessment for Machine Learning - From Test Data to Arguments
复制标题
DOI：
10.1007/978-3-030-00847-5_39
复制
发表时间：
2018-10
期刊：
影响因子：
--
通讯作者：
F. Ishikawa
中科院分区：
文献类型：
--
作者：
F. Ishikawa
已经积极努力使用机器学习（ML）技术来开发智能系统，例如，具有图像识别功能的驾驶支持系统。然而，ML组件的行为，例如，神经网络是从训练数据中归纳得出的，因此是不确定和不完美的。质量评估在很大程度上依赖于测试数据集或在大量可能性中尝试的内容，并受到这些内容的限制。鉴于这种独特的性质，我们提出了一个MLQ框架，用于评估ML组件和基于ML的系统的质量。我们引入概念来捕捉活动和证据的评估和支持建设的论点。
There have been active efforts to use machine learning (ML) techniques for the development of smart systems, e.g., driving support systems with image recognition. However, the behavior of ML components, e.g., neural networks, is inductively derived from training data and thus uncertain and imperfect. Quality assessment heavily depends on and is restricted by a test data set or what has been tried among an enormous number of possibilities. Given this unique nature, we propose a MLQ framework for assessing the quality of ML components and ML-based systems. We introduce concepts to capture activities and evidences for the assessment and support the construction of arguments.