Engineering Safety in Machine Learning

TL;DR

This paper discusses the importance of safety in machine learning systems, exploring definitions, challenges, and strategies for ensuring safety through design, interpretability, and human involvement, especially in high-stakes applications.

Contribution

It introduces a framework for understanding safety in machine learning, categorizes applications based on safety needs, and maps engineering safety strategies to ML practices.

Findings

Safety in ML involves risk, uncertainty, and harm considerations.

Different strategies like interpretability and human oversight can enhance safety.

Applications can be categorized into safety-critical and less critical types.

Abstract

Machine learning algorithms are increasingly influencing our decisions and interacting with us in all parts of our daily lives. Therefore, just like for power plants, highways, and myriad other engineered sociotechnical systems, we must consider the safety of systems involving machine learning. In this paper, we first discuss the definition of safety in terms of risk, epistemic uncertainty, and the harm incurred by unwanted outcomes. Then we examine dimensions, such as the choice of cost function and the appropriateness of minimizing the empirical average training cost, along which certain real-world applications may not be completely amenable to the foundational principle of modern statistical machine learning: empirical risk minimization. In particular, we note an emerging dichotomy of applications: ones in which safety is important and risk minimization is not the complete story (we name these Type A applications), and ones in which safety is not so critical and risk minimization is sufficient (we name these Type B applications). Finally, we discuss how four different strategies for achieving safety in engineering (inherently safe design, safety reserves, safe fail, and procedural safeguards) can be mapped to the machine learning context through interpretability and causality of predictive models, objectives beyond expected prediction accuracy, human involvement for labeling difficult or rare examples, and user experience design of software.