Dynamic Survival Analysis for Early Event Prediction
We show that using a model trained with a DSA likelihood outperforms models trained with a EEP likelihood on EEP metrics.
Mar 12, 2024
Improving Neural Additive Models with Bayesian Principles
Jan 1, 2024
On the Importance of Step-wise Embeddings for Heterogeneous Clinical Time-Series
Jan 1, 2023
Delphic Offline Reinforcement Learning under Nonidentifiable Hidden Confounding
Jan 1, 2023
Temporal Label Smoothing for Early Event Prediction
We show that by leveraging temporal dependencies in early event prediction in our objective, we can significantly improve performance over traditional MLE.
Sep 10, 2022
HiRID-ICU-Benchmark - A Comprehensive Machine Learning Benchmark on High-resolution ICU Data
We propose a large-scale benchmark for patient monitoring in the ICU, covering a wide range of tasks and providing a reproducible pipeline for data processing and evaluation.
Jul 3, 2021
Neighborhood Contrastive Learning Applied to Online Patient Monitoring
In this work, we propose a novel contrastive learning objective, Neighborhood Contrastive Learning (NCL), designed for data exhibiting a hierarchy. We apply it to online patient monitoring tasks.
Mar 5, 2021
UBS: A Dimension-Agnostic Metric for Concept Vector Interpretability Applied to Radiomics
Jan 1, 2019
Delphic Offline Reinforcement Learning under Nonidentifiable Hidden Confounding
Jan 1, 0001