Time-Series Foundation Model Embeddings for Remaining Useful Life Estimation
Summary
This paper introduces a lightweight approach for remaining useful life estimation using frozen embeddings from the Chronos-2 time-series foundation model combined with a simple regression head, achieving superior performance on industrial sensor data compared to baseline methods.
View Cached Full Text
Cached at: 06/11/26, 01:37 PM
Paper page - Time-Series Foundation Model Embeddings for Remaining Useful Life Estimation
Source: https://huggingface.co/papers/2606.11990
Abstract
A lightweight approach combining a frozen pretrained time-series foundation model with a simple regression head achieves superior RUL prediction performance compared to various baseline methods on industrial sensor data.
Remaining Useful Life (RUL) prediction is essential forindustrial predictive maintenance, yet many learning-based approaches rely on extensive feature engineering or large labeled datasets to train task-specific sequence models. In this work, we introduce a lightweight learning approach, in which we leverage a frozen pretrainedtime-series foundation model(TSFM) and combine it with a small regression head forRUL estimationfrommultivariate sensor streams. More specifically, we useChronos-2as afrozen backboneto extractcontext window featuresand train a lightweightregression neural networkfor RUL prediction. Experiments on real-world industrial sensor data from two device types show thatChronos-2features consistently improve over recurrent, convolutional, Transformer-based, and gradient-boosting baselines under the same preprocessing and evaluation protocol. We further analyze the impact of context length and find that performance improves significantly with longer histories, indicating that TSFM representation offer a practical and data-efficient alternative forRUL estimationin industrial settings.
View arXiv pageView PDFAdd to collection
Get this paper in your agent:
hf papers read 2606\.11990
Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash
Models citing this paper0
No model linking this paper
Cite arxiv.org/abs/2606.11990 in a model README.md to link it from this page.
Datasets citing this paper0
No dataset linking this paper
Cite arxiv.org/abs/2606.11990 in a dataset README.md to link it from this page.
Spaces citing this paper0
No Space linking this paper
Cite arxiv.org/abs/2606.11990 in a Space README.md to link it from this page.
Collections including this paper0
No Collection including this paper
Add this paper to acollectionto link it from this page.
Similar Articles
Are Time-Series Foundation Models Ready for E-Nose Data? An Empirical Assessment of Their Embeddings
This paper systematically evaluates time-series foundation models (TSFMs) such as Chronos-2 and MOMENT on electronic nose (E-Nose) data for gas identification and concentration prediction. It finds that fine-tuning is necessary and that fusing TSFM embeddings with specialized models can improve performance.
Scientific Machine Learning for Engine Health Management and Remaining Useful Life Prediction
This paper presents a multi-task scientific machine learning framework for turbine prognostics that jointly predicts engine health metrics and remaining useful life with quantified uncertainty, using a shared sequence encoder and task-specific heads.
A decoder-only foundation model for time-series forecasting
This article presents a research paper on Time-Series Foundation Model (TimeFM), a decoder-only model that achieves near-optimal zero-shot performance across diverse time-series datasets by adapting large language model techniques.
Assessing the Operational Viability of Foundation Models for Time Series Forecasting
This paper presents an applied evaluation of foundation models for time series forecasting compared to supervised approaches across four operational domains, and proposes a Complexity Router to selectively assign series to the optimal model class for balancing accuracy and inference cost.
A world model for the factory: predicting events across any machine, robot, or process from raw sensor streams
Forgis Labs presents a family of foundation models for time series sensor data in industrial settings, with five papers accepted to ICML 2026 workshops, enabling event prediction and natural language explanation from raw sensor streams.