SurvLoss: A New Survival Loss Function for Neural Networks to Process Censored Data

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Jun 27, 2024
Mahmoud Rahat Zahra Kharazian

Abstract

This paper presents SurvLoss, a novel asymmetric partial loss and error calculation function for survival analysis and regression, enabling the inclusion of censored samples. An observation in a dataset for which the complete information regarding an event of interest is not available is called censored. Censored samples are ubiquitous in the industry and play a crucial role in Prognostics and Health Management (PHM) by providing a realistic representation of data, improving the accuracy of analyses, and supporting better decision-making in various industries and the healthcare sector. The proposed approach can effectively equip the conventional regression loss functions such as Mean Absolute Error (MAE), Mean Squared Error (MSE), or Root Mean Squared Error (RMSE) with the ability to process censored samples. This can impact the field hugely by providing a more accessible usage of neural network models in survival analysis. The proposed survival loss incorporates censored samples by penalizing predictions outside the censoring region and skipping them otherwise. Then, it uses weighted averaging to aggregate the loss from censored samples with the loss from event samples. Unlike many other methods in the field, the proposed model distinguishes itself by avoiding superficial assumptions and exclusively relies on the available information, considering the entirety of the data.

We compared the proposed loss function with its baseline on two publicly available datasets. The first dataset, called C-MAPSS, is from NASA Turbofan Jet Engines simulation, and the second is a recently published real-world dataset from SCANIA trucks. The goal of both datasets is to predict the remaining useful life (RUL) of the machines.
The experimental results show that optimization algorithms for training deep neural networks like Adam can effectively utilize the proposed loss function to calculate gradients, update the model's weights, and reduce training and test errors. Moreover, the proposed model outperformed the baseline by taking advantage of the censored samples. The proposed loss function paves the way for the employment of advanced architectures of neural networks with bigger training sizes in survival analysis.

How to Cite

Rahat, M., & Kharazian, Z. . (2024). SurvLoss: A New Survival Loss Function for Neural Networks to Process Censored Data. PHM Society European Conference, 8(1), 7. https://doi.org/10.36001/phme.2024.v8i1.4052
Abstract 347 | PDF Downloads 183

##plugins.themes.bootstrap3.article.details##

Keywords

predictive Maintenance, Prognostics and Health Management, Survival Analysis, Remaining Useful Life prediction, Neural Networks, Loss Function, Censored Data, Label Uncertainty

References
@inproceedings{saxena2008damage,
title={Damage propagation modeling for aircraft engine run-to-failure simulation},
author={Saxena, Abhinav and Goebel, Kai and Simon, Don and Eklund, Neil},
booktitle={2008 international conference on prognostics and health management},
pages={1--9},
year={2008},
organization={IEEE}
}

@misc{Lindgren2024SCANIA,
doi={10.58141/1w9m-yz81},
language={sv},
publisher={Scania CV AB},
title={{SCANIA Component X Dataset: A Real-World Multivariate Time Series Dataset for Predictive Maintenance}},
url={https://doi.org/10.58141/1w9m-yz81},
author={Lindgren, Tony and Steinert, Olof and Andersson Reyna, Oskar and Kharazian, Zahra and Magnússon, Sindri},
date=2024,
year=2024,
}

@article{kharazian2024scania,
title={SCANIA Component X Dataset: A Real-World Multivariate Time Series Dataset for Predictive Maintenance},
author={Kharazian, Zahra and Lindgren, Tony and Magn{\'u}sson, Sindri and Steinert, Olof and Reyna, Oskar Andersson},
journal={arXiv preprint arXiv:2401.15199},
year={2024}
}

@inproceedings{rahat2023bridging,
title={Bridging the Gap: A Comparative Analysis of Regressive Remaining Useful Life Prediction and Survival Analysis Methods for Predictive Maintenance},
author={Rahat, Mahmoud and Kharazian, Zahra and Mashhadi, Peyman Sheikholharam and R{\"o}gnvaldsson, Thorsteinn and Choudhury, Shamik},
booktitle={PHM Society Asia-Pacific Conference},
volume={4},
number={1},
year={2023}
}

@inproceedings{rahat2020modeling,
title={Modeling turbocharger failures using markov process for predictive maintenance},
author={Rahat, Mahmoud and Pashami, Sepideh and Nowaczyk, S{\l}awomir and Kharazian, Zahra},
booktitle={30th European Safety and Reliability Conference (ESREL2020) \& 15th Probabilistic Safety Assessment and Management Conference (PSAM15), Venice, Italy, 1-5 November, 2020},
year={2020},
organization={European Safety and Reliability Association}
}

@inproceedings{rahat2022domain,
title={Domain adaptation in predicting turbocharger failures using vehicle’s sensor measurements},
author={Rahat, Mahmoud, et al},
booktitle={Phm society european conference},
volume={7},
number={1},
pages={432--439},
year={2022}
}


@inproceedings{altarabichi2020stacking,
title={Stacking ensembles of heterogenous classifiers for fault detection in evolving environments},
author={Altarabichi, Mohammed Ghaith, et al},
booktitle={30th European Safety and Reliability Conference, ESREL 2020 and 15th Probabilistic Safety Assessment and Management Conference, PSAM15 2020, Venice, Italy, 1-5 November, 2020},
pages={1068--1068},
year={2020},
organization={Research Publishing Services}
}


@inproceedings{revanur2020embeddings,
title={Embeddings based parallel stacked autoencoder approach for dimensionality reduction and predictive maintenance of vehicles},
author={Revanur, Vandan, et al},
booktitle={IoT Streams for Data-Driven Predictive Maintenance and IoT, Edge, and Mobile for Embedded Machine Learning: Second International Workshop, IoT Streams 2020, and First International Workshop, ITEM 2020, Co-located with ECML/PKDD 2020, Ghent, Belgium, September 14-18, 2020, Revised Selected Papers 2},
pages={127--141},
year={2020},
organization={Springer}
}

@inproceedings{karlsson2023baseline,
title={Baseline Selection for Integrated Gradients in Predictive Maintenance of Volvo Trucks’ Turbocharger},
author={Karlsson, Nellie, et al},
booktitle={VEHICULAR 2023-IARIA},
year={2023}
}

@article{wang2019machine,
title={Machine learning for survival analysis: A survey},
author={Wang, Ping and Li, Yan and Reddy, Chandan K},
journal={ACM Computing Surveys (CSUR)},
volume={51},
number={6},
pages={1--36},
year={2019},
publisher={ACM New York, NY, USA}
}

@article{cox1972regression,
title={Regression models and life-tables},
author={Cox, David R},
journal={Journal of the Royal Statistical Society: Series B (Methodological)},
volume={34},
number={2},
pages={187--202},
year={1972},
publisher={Wiley Online Library}
}

@article{harrell1982evaluating,
title={Evaluating the yield of medical tests},
author={Harrell, Frank E and Califf, Robert M and Pryor, David B and Lee, Kerry L and Rosati, Robert A},
journal={Jama},
volume={247},
number={18},
pages={2543--2546},
year={1982},
publisher={American Medical Association}
}

@article{katzman2018deepsurv,
title={DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network},
author={Katzman, Jared L and Shaham, Uri and Cloninger, Alexander and Bates, Jonathan and Jiang, Tingting and Kluger, Yuval},
journal={BMC medical research methodology},
volume={18},
pages={1--12},
year={2018},
publisher={Springer}
}

@article{kvamme2019time,
title={Time-to-event prediction with neural networks and Cox regression},
author={Kvamme, H{\aa}vard and Borgan, {\O}rnulf and Scheel, Ida},
journal={Journal of machine learning research},
volume={20},
number={129},
pages={1--30},
year={2019}
}

@inproceedings{chen2020deep,
title={Deep kernel survival analysis and subject-specific survival time prediction intervals},
author={Chen, George H},
booktitle={Machine Learning for Healthcare Conference},
pages={537--565},
year={2020},
organization={PMLR}
}

@article{ishwaran2008random,
title={Random survival forests},
author={Ishwaran, Hemant and Kogalur, Udaya B and Blackstone, Eugene H and Lauer, Michael S},
year={2008}
}

@article{hartman2023pitfalls,
title={Pitfalls of the concordance index for survival outcomes},
author={Hartman, Nicholas and Kim, Sehee and He, Kevin and Kalbfleisch, John D},
journal={Statistics in Medicine},
volume={42},
number={13},
pages={2179--2190},
year={2023},
publisher={Wiley Online Library}
}

@article{alabdallah2024concordance,
title={The Concordance Index decomposition: A measure for a deeper understanding of survival prediction models},
author={Alabdallah, Abdallah and Ohlsson, Mattias and Pashami, Sepideh and R{\"o}gnvaldsson, Thorsteinn},
journal={Artificial Intelligence in Medicine},
volume={148},
pages={102781},
year={2024},
publisher={Elsevier}
}

@book{kleinbaum1996survival,
title={Survival analysis a self-learning text},
author={Kleinbaum, David G and Klein, Mitchel},
year={1996},
publisher={Springer}
}
Section
Technical Papers