Learning from data with structured missingness
Learning from data with structured missingness
复制标题
DOI：
10.1038/s42256-022-00596-z
复制
发表时间：
2023-01-01
影响因子：
23.8
通讯作者：
MacArthur，Ben D.
中科院分区：
文献类型：
--
作者：
Mitra，Robin;McGough，Sarah F.;MacArthur，Ben D.
丢失数据是许多机器学习任务中不可避免的复杂问题。当数据“随机缺失”时，存在一系列工具和技术来处理这个问题。然而，随着机器学习研究变得越来越雄心勃勃，并试图从越来越大的异构数据中学习，越来越多地遇到这样的问题，即缺失值显式或隐式地表现出关联或结构。这种“结构化缺失”带来了一系列尚未得到系统解决的挑战，并对大规模机器学习构成了根本性的障碍。在这里，我们概述了当前的文献，并提出了一系列从结构化缺失数据中学习的重大挑战。
Missing data are an unavoidable complication in many machine learning tasks. When data are ‘missing at random’ there exist a range of tools and techniques to deal with the issue. However, as machine learning studies become more ambitious, and seek to learn from ever-larger volumes of heterogeneous data, an increasingly encountered problem arises in which missing values exhibit an association or structure, either explicitly or implicitly. Such ‘structured missingness’ raises a range of challenges that have not yet been systematically addressed, and presents a fundamental hindrance to machine learning at scale. Here we outline the current literature and propose a set of grand challenges in learning from data with structured missingness.