arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2002.05193cs.CYcs.LGecon.EMmath.STstat.MLstat.TH
收藏

机器学习中的限制层次

A Hierarchy of Limitations in Machine Learning

Berkman Klein Center for Internet & Society at Harvard University(哈佛大学伯克曼·克莱因互联网与社会中心)

Momin M. Malik

摘要:

“所有的模型都是错误的,但有些是有用的,”乔治E。P. Box(1979).机器学习一直专注于概率模型在社会系统预测中的有用性,但直到现在才开始掌握这些模型错误的方式以及这些缺陷的后果。本文试图对应用于社会时机器学习中模型的具体概念、程序和统计限制进行全面、结构化的概述。机器学习建模者自己可以使用所描述的层次结构来识别可能的故障点,并考虑如何解决这些问题,而机器学习模型的消费者可以知道在面临关于是否、在何处以及如何应用机器学习的决策时应该质疑什么。局限性从量化本身固有的承诺,到显示未建模的依赖关系如何导致交叉验证作为评估模型性能的一种方式过于乐观。

英文摘要:

"All models are wrong, but some are useful", wrote George E. P. Box (1979). Machine learning has focused on the usefulness of probability models for prediction in social systems, but is only now coming to grips with the ways in which these models are wrong---and the consequences of those shortcomings. This paper attempts a comprehensive, structured overview of the specific conceptual, procedural, and statistical limitations of models in machine learning when applied to society. Machine learning modelers themselves can use the described hierarchy to identify possible failure points and think through how to address them, and consumers of machine learning models can know what to question when confronted with the decision about if, where, and how to apply machine learning. The limitations go from commitments inherent in quantification itself, through to showing how unmodeled dependencies can lead to cross-validation being overly optimistic as a way of assessing model performance.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑