ICU-Sepsis: A Benchmark MDP Built from Real Medical Data

Kartik Choudhary, Dhawal Gupta, Philip S. Thomas
6/9/2024
cs.LG

No official code found. Best implementation: icu-sepsis/icu-sepsis (43 stars, MIT, last commit 2024-10-23). 1 more below.

Abstract

We present ICU-Sepsis, an environment that can be used in benchmarks for evaluating reinforcement learning (RL) algorithms. Sepsis management is a complex task that has been an important topic in applied RL research in recent years. Therefore, MDPs that model sepsis management can serve as part of a benchmark to evaluate RL algorithms on a challenging real-world problem. However, creating usable MDPs that simulate sepsis care in the ICU remains a challenge due to the complexities involved in acquiring and processing patient data. ICU-Sepsis is a lightweight environment that models personalized care of sepsis patients in the ICU. The environment is a tabular MDP that is widely compatible and is challenging even for state-of-the-art RL algorithms, making it a valuable tool for benchmarking their performance. However, we emphasize that while ICU-Sepsis provides a standardized environment for evaluating RL algorithms, it should not be used to draw conclusions that guide medical practice.

View on arXivView PDF

Code Implementations(2)

icu-sepsis/icu-sepsisCommunity100%
MIT

ICU-Sepsis is a lightweight, yet challenging RL environment that models the treatment of sepsis in the ICU.

432PythonFeb 29, 20241 years agoMIT
applied-reinforcement-learningreinforcement-learningreinforcement-learning-environments
Dhawgupta/choudhary2024icuCommunity54%

Repository which contains implementation of baselines algorithms including PPO, DQN and SAC for the ICU Sepsis benchmark (https://github.com/icu-sepsis/icu-sepsis), introduced in "ICU-Sepsis: A Benchmark MDP Built from Real Medical Data", accepted in RLC 2024.

31PythonOct 21, 20242 years ago

Similar Papers

Formally Verifying and Explaining Sepsis Treatment Policies with COOL-MC

87% match2026

How Consistent are Clinicians? Evaluating the Predictability of Sepsis Disease Progression with Dynamics Models

85% match1 repo292024

Reinforcement Learning in Dynamic Treatment Regimes Needs Critical Reexamination

85% match1 repo142024

MORE-CLEAR: Multimodal Offline Reinforcement learning for Clinical notes Leveraged Enhanced State Representation

84% match2025

Attention-Based Offline Reinforcement Learning and Clustering for Interpretable Sepsis Treatment

83% match2026

Exploring Time-Step Size in Reinforcement Learning for Sepsis Treatment

83% match2025

Cite this paper

@article{choudhary2024icusepsis,
  title  = {ICU-Sepsis: A Benchmark MDP Built from Real Medical Data},
  author = {Kartik Choudhary and Dhawal Gupta and Philip S. Thomas},
  year   = {2024},
  eprint = {2406.05646},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2406.05646v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2406.05646.svg)](https://codeofpaper.com/paper/2406.05646)

Discussion