Interpretable Predictive Maintenance for Hard Drives

TL;DR

This paper explores interpretable machine learning models for predicting hard drive failures, balancing high predictive accuracy with human-understandable insights, even with limited historical data.

Contribution

It demonstrates that interpretable models can effectively predict drive failures and provide meaningful insights, addressing the black-box nature of traditional approaches.

Findings

Interpretable models maintain high predictive performance.

Insights are meaningful for understanding failure mechanisms.

Models are effective even with limited historical data.

Abstract

Existing machine learning approaches for data-driven predictive maintenance are usually black boxes that claim high predictive power yet cannot be understood by humans. This limits the ability of humans to use these models to derive insights and understanding of the underlying failure mechanisms, and also limits the degree of confidence that can be placed in such a system to perform well on future data. We consider the task of predicting hard drive failure in a data center using recent algorithms for interpretable machine learning. We demonstrate that these methods provide meaningful insights about short- and long-term drive health, while also maintaining high predictive performance. We also show that these analyses still deliver useful insights even when limited historical data is available, enabling their use in situations where data collection has only recently begun.