Trajectory Flow Matching with Applications to Clinical Time Series Modelling
Authors: Xi (Nicole) Zhang, Yuan Pu, Yuki Kawamura, Andrew Loza, Yoshua Bengio, Dennis Shung, Alexander Tong
NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We empirically demonstrate the utility of our method in clinical applications where hemodynamic trajectories are critical for ongoing dynamic monitoring and care. We applied our method to the following longitudinal electronic health record datasets: medical intensive care unit (MICU) data of patients with sepsis, ICU patients at risk for cardiac arrest, Emergency Department (ED) data of patients with acute gastrointestinal bleeding, and MICU data of patients with acute gastrointestinal bleeding. In this section we empirically evaluate the performance of the trajectory flow matching objective in terms of time series modeling error, but also uncertainty quantification. |
| Researcher Affiliation | Academia | 1Mc Gill University, 2Mila Quebec AI Institute, 3Yale School of Medicine 4School of Clinical Medicine, University of Cambridge, 5Université de Montréal, 6CIFAR Fellow |
| Pseudocode | Yes | Algorithm 1 General Trajectory Flow Matching |
| Open Source Code | Yes | Code available at: https://github.com/n Zhangx/Trajectory Flow Matching |
| Open Datasets | Yes | ICU Sepsis: a subset of the e ICU Collaborative Research Database v2.0 [Pollard et al., 2019] of patients admitted with sepsis as the primary diagnosis. ICU GIB: a subset of the Medical Information Mart for Intensive Care III [Johnson et al., 2016] of patients with gastrointestinal bleeding as the primary diagnosis. The publicly available datasets from MIMIC-III and e ICU databases are properly credited, respected, mentioned and used under the Physio Net Credentialed Health Data License Version 1.5.0. |
| Dataset Splits | Yes | The ICU Sepsis Dataset was created by subsetting the e ICU Database for 3362 patients with sepsis as the primary admission diagnosis (2689 patients in training set, 336 in validation set, and 337 in test set). |
| Hardware Specification | Yes | Experiments were run on a computing cluster with a heterogenous cluster of NVIDIA RTX8000, V100, A40, and A100 GPUs for approximately 24,000 GPU hours. |
| Software Dependencies | No | The paper mentions 'Adam optimizer' and 'GRU units' and 'XGBoost model', but does not specify version numbers for these or any other software dependencies. |
| Experiment Setup | Yes | All the models for clinical data experiments are trained with Adam optimizer. A maximum training time and epochs are set to 48 hours and 300, with early stopping (patience=3) monitoring validation loss. All metrics reported were ran with 5 seeds (0,1,2,3,4) to ensure it is reproducible. TFM, TFM-ODE, and ablations The TFM models were trained with learning rate 1 10 6 and had σ of 0.1. The complete models have hidden size of 256 and memory of 3, while ablation study with a hidden size of 64 and/or no memory was performed (Table 4). The noise parameter for the SDE implementation was set to 0.1 for ablations without Luncertainty. The hyperparameters σ = 0.1 and memory=3 for full models were selected through experiments with different values of σ and memory (Figure 5 and 6). |