Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks
Julie Alberge (PhD Student, inria Saclay, Soda team)
| When |
Jun 17, 2025
from 01:00 to 02:00 |
|---|---|
| Where | M7 101 |
| Attendees |
Julie Alberge |
| Add event to calendar |
vCal
iCal
|
Julie Alberge
Title: Survival Models: Proper Scoring Rule and Stochastic Optimization with Competing Risks
Abstract : When dealing with right-censored data, where some outcomes are missing due to a limited observation period, survival 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, an area known as competing risks.
Here, we present a new strictly proper separable scoring rule to handle competing risks. Then, we propose SurvivalBoost,
a gradient boosting trees implemented with the previous loss that outperforms 12 state-of-the-art models across several
metrics, both in competing risks and survival settings, and with faster computation times compared to existing methods.
on predicting the time until an event of interest occurs. Multiple classes of outcomes lead to a classification variant: predicting
the most likely event, an area known as competing risks.
Here, we present a new strictly proper separable scoring rule to handle competing risks. Then, we propose SurvivalBoost,
a gradient boosting trees implemented with the previous loss that outperforms 12 state-of-the-art models across several
metrics, both in competing risks and survival settings, and with faster computation times compared to existing methods.
Website: https://jualberge.github.io/
In Room M7 101, 1st floor, Monod campus, ENSL.