文章基本信息

标题：Steampunk Machine Learning:Victorian contrivances for modern data science
本地全文：下载
作者：Terence Kelly
期刊名称：ACM Queue (Online): tomorrow's computing today
电子版ISSN：1542-7749
出版年度：2021
卷号：19
期号：6
页码：1-13
语种：English
出版社：Association for Computing Machinery
摘要：Like all engineers, programmers and datacenter operators employ models to describe and predict the behavior of complex production systems. For example, a performance expert might define a parameterized equation that maps a computer system's workload to its application-level performance under ordinary conditions. Fitting such a model to data—setting the free parameters so that the equation agrees closely with measurements of the running system—is straightforward in principle.