Calibration and Uncertainty in Neural Time-to-Event Modeling (IEEE TNNLS 2020)

This repository contains links to TensorFlow code for replicating experiments in our journal article Calibration and Uncertainty in Neural Time-to-Event Modeling to appear as a Special Issue: Robust Learning of Spatio-Temporal Point Processes: Modeling, Algorithm, and Applications at IEEE TNNLS

@article{chapfuwa2020calibration,
  title={Calibration and Uncertainty in Neural Time-to-Event Modeling},
  author={Chapfuwa, Paidamoyo and Tao, Chenyang and Li, Chunyuan and Khan, Irfan and Chandross, Karen J and Pencina, Michael J and Carin, Lawrence and Henao, Ricardo},
  journal={IEEE Transactions on Neural Networks and Learning Systems},
  year={2020},
  publisher={IEEE}
}

Calibration in Time-to-Event Models

• We propose a new estimator that can be used to visually assess the calibration (accounting for model uncertainty) of estimated event times from different models relative to the ground truth
• Run the Calibration.ipynb to generate calibration results

Proposed Models

We propose the following models implemented here:

• An AFT plus ranking baseline DRAFT
• An adversarial nonparametric model DATE
• We also consider replacing the discriminator of the adversarial nonparametric model with a survival-function matching estimator SFM that accounts for model calibration

Acknowledgments

This work leverages the accuracy objective from DATE. Contact Paidamoyo for issues relevant to this project.