Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks

Julie Alberge, Vincent Maladiere, Olivier Grisel, Judith Abécassis, Gael Varoquaux
Proceedings of The 28th International Conference on Artificial Intelligence and Statistics, PMLR 258:3619-3627, 2025.

Abstract

When dealing with right-censored data, where some outcomes are missing due to a limited observation period, survival analysis —known as \emph{time-to-event analysis}— focuses on predicting the time until an event of interest occurs. Multiple classes of outcomes lead to a classification variant: predicting the most likely event, a less explored area known as \emph{competing risks}. Classic competing risks models couple architecture and loss, limiting scalability. To address these issues, we design a strictly proper censoring-adjusted separable scoring rule, allowing optimization on a subset of the data as each observation is evaluated independently. The loss estimates outcome probabilities and enables stochastic optimization for competing risks, which we use for efficient gradient boosting trees. \textbf{SurvivalBoost} not only outperforms 12 state-of-the-art models across several metrics on 4 real-life datasets, both in competing risks and survival settings, but also provides great calibration, the ability to predict across any time horizon, and computation times faster than existing methods.

Cite this Paper

BibTeX
@InProceedings{pmlr-v258-alberge25a, title = {Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks}, author = {Alberge, Julie and Maladiere, Vincent and Grisel, Olivier and Ab{\'e}cassis, Judith and Varoquaux, Gael}, booktitle = {Proceedings of The 28th International Conference on Artificial Intelligence and Statistics}, pages = {3619--3627}, year = {2025}, editor = {Li, Yingzhen and Mandt, Stephan and Agrawal, Shipra and Khan, Emtiyaz}, volume = {258}, series = {Proceedings of Machine Learning Research}, month = {03--05 May}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v258/main/assets/alberge25a/alberge25a.pdf}, url = {https://proceedings.mlr.press/v258/alberge25a.html}, abstract = {When dealing with right-censored data, where some outcomes are missing due to a limited observation period, survival analysis —known as \emph{time-to-event analysis}— focuses on predicting the time until an event of interest occurs. Multiple classes of outcomes lead to a classification variant: predicting the most likely event, a less explored area known as \emph{competing risks}. Classic competing risks models couple architecture and loss, limiting scalability. To address these issues, we design a strictly proper censoring-adjusted separable scoring rule, allowing optimization on a subset of the data as each observation is evaluated independently. The loss estimates outcome probabilities and enables stochastic optimization for competing risks, which we use for efficient gradient boosting trees. \textbf{SurvivalBoost} not only outperforms 12 state-of-the-art models across several metrics on 4 real-life datasets, both in competing risks and survival settings, but also provides great calibration, the ability to predict across any time horizon, and computation times faster than existing methods.} }
Endnote
%0 Conference Paper %T Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks %A Julie Alberge %A Vincent Maladiere %A Olivier Grisel %A Judith Abécassis %A Gael Varoquaux %B Proceedings of The 28th International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2025 %E Yingzhen Li %E Stephan Mandt %E Shipra Agrawal %E Emtiyaz Khan %F pmlr-v258-alberge25a %I PMLR %P 3619--3627 %U https://proceedings.mlr.press/v258/alberge25a.html %V 258 %X When dealing with right-censored data, where some outcomes are missing due to a limited observation period, survival analysis —known as \emph{time-to-event analysis}— focuses on predicting the time until an event of interest occurs. Multiple classes of outcomes lead to a classification variant: predicting the most likely event, a less explored area known as \emph{competing risks}. Classic competing risks models couple architecture and loss, limiting scalability. To address these issues, we design a strictly proper censoring-adjusted separable scoring rule, allowing optimization on a subset of the data as each observation is evaluated independently. The loss estimates outcome probabilities and enables stochastic optimization for competing risks, which we use for efficient gradient boosting trees. \textbf{SurvivalBoost} not only outperforms 12 state-of-the-art models across several metrics on 4 real-life datasets, both in competing risks and survival settings, but also provides great calibration, the ability to predict across any time horizon, and computation times faster than existing methods.
APA
Alberge, J., Maladiere, V., Grisel, O., Abécassis, J. & Varoquaux, G.. (2025). Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks. Proceedings of The 28th International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 258:3619-3627 Available from https://proceedings.mlr.press/v258/alberge25a.html.

Related Material