Evaluating deep learning models for fault diagnosis of a rotating machinery with epistemic and aleatoric uncertainty
arXiv:2412.18980v2 Announce Type: replace Abstract: Uncertainty-aware deep learning (DL) models recently gained attention in fault diagnosis as a way to promote the reliable detection of faults when out-of-distribution (OOD) data arise from unseen faults (epistemic uncertainty) or the presence of noise (aleatoric uncertainty). In this paper, we present the first comprehensive comparative study of state-of-the-art uncertainty-aware DL architectures for fault diagnosis in rotating machinery, where different scenarios affected by epistemic uncertainty and different types of aleatoric uncertainty
The increasing adoption of deep learning in critical industrial applications necessitates robust methods for understanding and mitigating uncertainty, making this research timely.
Reliable fault diagnosis in machinery is crucial for operational efficiency, safety, and predictive maintenance, with AI models needing to account for real-world unpredictability.
This research advances the practical application of uncertainty-aware deep learning, improving the trustworthiness and deployment of AI in industrial settings.
- · Industrial automation sector
- · Maintenance and reliability engineering
- · AI/ML model developers
- · Manufacturing companies
- · Companies relying on traditional, less accurate fault detection methods
Increased operational uptime and reduced unexpected failures in industrial machinery.
Higher confidence in deploying AI systems for critical functions, accelerating AI adoption in production environments.
Potentially enables new insurance models for industrial assets based on AI-driven predictive reliability.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG