Towards Unified and Data-Efficient Prognostics and Health Management with Tabular Foundation Models

2026-06-03

AI Summary

This work addresses the challenges posed by fragmented, partially observed, and sparsely labeled data in industrial Prognostics and Health Management (PHM), which hinder the effectiveness of conventional supervised learning approaches. It introduces, for the first time, a tabular foundation model to the PHM domain by transforming raw time-series signals into a tabular format that preserves temporal context, enabling unified multi-task modeling for both fault diagnosis and remaining useful life prediction through in-context learning. Experimental results demonstrate that the proposed method achieves state-of-the-art average performance across diverse PHM tasks and significantly outperforms sequence models, Transformers, and gradient-boosted trees—particularly in low-data regimes—thereby validating the efficiency, generalization capability, and versatility of tabular foundation models for heterogeneous PHM problems.

Abstract

Data-driven Prognostics and Health Management (PHM) uses time-varying condition-monitoring data to diagnose system states and estimate remaining useful life in engineered assets. These tasks are central to maintenance planning, but industrial PHM data are often fragmented, partially observed, and poorly labeled, which hinders supervised learning. Foundation models offer a route toward reusable predictive systems, yet most time-series foundation models are designed for forecasting and assume long, coherent, regularly sampled sequences. To address this gap, we propose a framework for applying Tabular Foundation Models to industrial time series using in-context learning, and we evaluate them on a variety of PHM tasks. By converting raw unit-level signals into tabular rows, we show that these models perform well across multiple tasks - including prognostics, and diagnostics - and are highly data efficient. We compare them directly with sequence models, transformer baselines, and gradient-boosted trees under a common evaluation protocol. The results indicate that tabular foundation models achieve the best average ranks across prognostic and diagnostic tasks. Our findings further show that PFN-based models are competitive in low-data regimes, that temporal context can be preserved in the tabular representation, and that performance depends on representative context construction under subsampling. These results demonstrate that tabular foundation models provide a practical and general interface for heterogeneous PHM problems.

Problem

• Prognostics and Health Management
• Tabular Foundation Models
• Data Efficiency
• Industrial Time Series
• In-context Learning

Innovation

• Tabular Foundation Models
• Prognostics and Health Management
• In-context Learning
• Data Efficiency
• Industrial Time Series