Deep learning and lazy learner improve early sepsis detection.
problem Detecting sepsis early in high-resolution ICU records.
method Deep learning model with temporal convolutional network and Gaussian Process Adapter; lazy learner with dynamic time warping.
result Improves sepsis detection from 0.25 to 0.40 AUPRC 7 hours before onset.
We study multiple rule-based and machine learning (ML) models for sepsis detection. We report the first neural network detection and prediction results on three categories of sepsis. We have used the retrospective Medical Information Mart for Intensive Care (MIMIC)-III dataset, restricted to intensive care unit (ICU) p…
Deep learning detects sepsis early, improving patient outcomes.
problem Improving timeliness of sepsis detection in various healthcare settings.
method Deep learning model on electronic health records, combining CNN and LSTM networks.
result Model performance ranges from 0.856 to 0.756 AUROC, detecting sepsis up to 24 hours in advance.
Paper uses ensemblers to predict sepsis early from patient records.
problem Early detection of sepsis in patients.
method Imputation and weak ensembler technique applied to 40k patient records.
result Model achieved 93.45% accuracy and 0.271 utility score.
Sepsis is a poorly understood and potentially life-threatening complication that can occur as a result of infection. Early detection and treatment improves patient outcomes, and as such it poses an important challenge in medicine. In this work, we develop a flexible classifier that leverages streaming lab results, vita…
Model predicts sepsis early with high accuracy.
problem Early detection of sepsis to reduce mortality.
method MGP-AttTCN: Gaussian Process and attention-based deep learning.
result Model outperforms state-of-the-art with AUROC 0.660 and AUPR 0.483.
Although timely sepsis diagnosis and prompt interventions in Intensive Care Unit (ICU) patients are associated with reduced mortality, early clinical recognition is frequently impeded by non-specific signs of infection and failure to detect signs of sepsis-induced organ dysfunction in a constellation of dynamically cha…
Study identifies sepsis subpopulations for better ICU prediction.
problem Lack of stratified sepsis subpopulations in ICU EHR predictions.
method Stratified sepsis subpopulations identified using machine learning.
result Machine learning models perform better with stratified subpopulations.
Sepsis is a dangerous condition that is a leading cause of patient mortality. Treating sepsis is highly challenging, because individual patients respond very differently to medical interventions and there is no universally agreed-upon treatment for sepsis. In this work, we explore the use of continuous state-space mode…
Characterizing a patient's progression through stages of sepsis is critical for enabling risk stratification and adaptive, personalized treatment. However, commonly used sepsis diagnostic criteria fail to account for significant underlying heterogeneity, both between patients as well as over time in a single patient. W…
New dataset from clinicians improves sepsis prediction models.
problem Circularity in previous sepsis prediction models.
method Developed an independent dataset from clinical judgments, avoiding circularity.
result Achieved state-of-the-art AUROC scores.
We present a scalable end-to-end classifier that uses streaming physiological and medication data to accurately predict the onset of sepsis, a life-threatening complication from infections that has high mortality and morbidity. Our proposed framework models the multivariate trajectories of continuous-valued physiologic…
DeepAISE predicts sepsis onset with high accuracy and low false alarms.
problem Early prediction of sepsis in ICU patients to improve clinical situational awareness.
method Recurrent neural survival model that combines clinical criteria and treatment policies.
result DeepAISE produces the most accurate predictions (AUC=0.90 and 0.87) and lowest false alarm rates (FAR=0.20 and 0.26) compared to baseline models.
EHR-MPC optimizes sepsis treatment using digital twins and inference-time control.
problem Optimal sepsis treatment policies are contested and difficult to adapt during inference.
method EHR-MPC decouples learning patient dynamics from treatment optimization, enabling inference-time control over learned digital twins.
result EHR-MPC achieves comparable off-policy performance and improved simulation performance compared to RL baselines.
Simulates sepsis treatment decisions using a world model approach.
problem Predicting optimal sepsis treatment actions based on noisy EHR data.
method Uses a Variational Auto-Encoder and Mixture Density Network (MDN-RNN) to model sepsis patient trajectories.
result Simulator learns from MIMIC dataset to predict patient states.
Paper uses a mix of deep and kernel learning to personalize sepsis treatment.
problem Managing sepsis in ICU patients due to individual variability.
method A mixture-of-experts framework combining kernel-based and deep reinforcement learning.
result The mixture-based approach outperforms individual methods on a large sepsis patient cohort.
Study identifies subphenotypes of pediatric sepsis to improve ML predictive performance.
problem Enhance machine learning predictive performance in pediatric sepsis.
method Latent profile analysis of clinical data to identify subphenotypes, followed by ML experiments.
result Improved predictive performance of ML models targeting specific subphenotypes of pediatric sepsis.
Machine learning models trained on indirect data labels can fail on real-world examples.
problem Validity issues in machine learning when target labels are indirectly defined.
method Identification of problematic datasets and models using a general procedure.
result Machine learning models trained on indirect data labels will fail on real-world examples.
New method identifies sepsis-related patient features in EMR data.
problem Identify sepsis-related patient features in EMR data.
method Linear multivariate Hawkes process model with ReLU link function, coupled with gradient-based method.
result Identifies several interpretable GC chains that precede sepsis.
Predict sepsis early from EHR data with aggregated clinical events.
problem Predict sepsis from clinical data in EHR with temporal interactions.
method Aggregates heterogeneous clinical events, captures temporal interactions with LSTM.
result Achieved high utility score (0.321) in PhysioNet/Computing in Cardiology Challenge 2019.
Study evaluates how framing affects machine learning models for sepsis prediction.
problem Understanding and reporting framing is crucial for AI technology success.
method Four different approaches applied to AI models of sepsis prediction.
result On-clinical-demand framing showed the lowest missing values and best temporal dependencies.
HyDaP clusters mixed-type data efficiently.
problem Clustering data with mixed types (continuous and categorical).
method Two-step hybrid approach: density-based and partition-based for continuous variables, partition-based for mixed data.
result HyDaP outperforms existing methods in clustering electronic health records.
Paper detects bias in AI medical models using CART.
problem Ensuring fairness in AI medical decision support systems.
method Uses Classification and Regression Trees (CART) algorithm to identify bias.
result Validated the CART approach in both synthetic and real-world data.
New methods for evaluating and optimizing policies in offline RL with unobserved confounders.
problem Evaluating and optimizing policies in the presence of unobserved confounders.
method Characterized settings and algorithms for consistent value estimates and lower bounds, with sample complexity guarantees.
result Proved local convergence guarantees for offline policy improvement.
Study investigates how machine learning models degrade over time, leading to patient safety issues.
problem Overtime degradation of machine learning models in clinical settings.
method Used MIMIC-IV dataset to train models replicating commercial approaches, observing and analyzing degradation over a decade.
result An RNN model built on Epic features degrades from 0.729 AUC to 0.525 AUC over a decade, highlighting technical and clinical drift as root causes.
New algorithm learns expert reward structures from batch data.
problem Learning expert reward structures from batch data without dynamics models.
method Deep Successor Feature Networks (DSFN) and transition-regularized imitation network.
result Superior performance on control benchmarks and sepsis management.
Sepsis is a life-threatening disease and one of the major causes of death in hospitals. Imaging of microcirculatory dysfunction is a promising approach for automated diagnosis of sepsis. We report a machine learning classifier capable of distinguishing non-septic and septic images from dark field microcirculation video…
DCEM algorithm reduces bias in machine learning models trained on selective labels.
problem Bias in machine learning models trained on selective labels.
method Disparate Censorship Expectation-Maximization (DCEM) algorithm.
result DCEM improves bias mitigation without sacrificing discriminative performance.
Estimates individualized treatment effects using shared RBF-net neurons.
problem Identifying differential treatment effects based on covariates.
method Non-parametric radial basis function (RBF)-nets with shared hidden neurons in a Bayesian framework.
result Demonstrated through simulations and real data, the method identifies interesting treatment effects.
A TCL framework improves causal effect estimation in limited data.
problem Improving causal effect estimation accuracy in limited data.
method Transfer Learning (TCL) with ℓ1 regularization for nuisance models.
result Non-asymptotic recovery guarantees for exttt{ℓ1-TCL} in high-dimensional settings. Proposes a method to estimate personalized treatments from high-dimensional data.
problem Estimating individualized treatment regimes (ITRs) from high-dimensional covariates.
method Directly targets the contrast between potential outcomes, using dimension-reduced outcome-weighted learning.
result Achieves universal consistency, converging to the Bayes risk under mild conditions.
Study analyzes sensitivity of RL algorithm for ICU hemodynamic management.
problem Evaluating safety and reliability of RL in clinical settings.
method Sensitivity analysis of Duel-DDQN on ICU sepsis patients.
result RL policies are sensitive to various implementation factors.
Causal thinking improves healthcare decisions from EHRs.
problem Shortcuts in data lead to biased healthcare decisions.
method Step-by-step framework for valid decision making from EHRs.
result Valid decision making requires careful analysis of EHR data.
New method identifies proxies for causal effects on multiple outcomes.
problem Estimating causal effects in scenarios with multiple outcomes and treatments.
method Causal discovery method leveraging multiple outcomes as proxies for each treatment effect.
result Parallel studies of multiple outcomes can assist in causal identification.
New method estimates policy performance under unobserved confounding.
problem Estimating policy performance when decisions depend on unobserved variables.
method Developed worst-case bounds for robust OPE under unobserved confounding.
result Efficient procedure for computing worst-case bounds, proving statistical consistency.
PRL improves off-policy evaluation in partially observed MDPs.
problem Confounding and bias in offline reinforcement learning with unobserved state factors.
method Extends proximal causal inference to POMDPs, identifying and estimating target policy value.
result Semiparametrically efficient estimators for PRL in partially observed MDPs.
Bayesian model for multi-environment prediction with latent variable changes.
problem Prediction in environments with changing latent variable distributions.
method Bayesian model with empirical Bayes prior and amortized variational algorithm.
result Method outperforms previous approaches in new environments.
Framework improves clinical timeline reconstruction from text and tables.
problem Temporal precision and event timing in clinical narratives and EHRs.
method Retrieval-augmented multimodal alignment framework.
result Consistently improves absolute timestamp accuracy and temporal concordance.
The lack of interpretability remains a key barrier to the adoption of deep models in many applications. In this work, we explicitly regularize deep models so human users might step through the process behind their predictions in little time. Specifically, we train deep time-series models so their class-probability pred…
Bayesian optimization improves classifier selection for acute infection and mortality.
problem Improving accuracy of acute infection and mortality prediction.
method Comparison of hyperparameter optimization methods (grid search, random sampling, Bayesian optimization).
result Bayesian optimization outperforms grid search or random sampling for in-hospital mortality classifiers.
Estimating the causal effects of an intervention from high-dimensional observational data is difficult due to the presence of confounding. The task is often complicated by the fact that we may have a systematic missingness in our data at test time. Our approach uses the information bottleneck to perform a low-dimension…
Procedure for evaluating RL policies in risky settings.
problem Identifying risky episodes in RL policy applications.
method Gumbel-Max SCMs for generating counterfactual trajectories.
result Identifies episodes with dramatic counterfactual reward differences.
A new method explains RNNs by decision lists over skipgrams, improving explanation fidelity and interpretability.
problem Lack of understanding how input segments combine to form patterns in neural network outputs.
method Proposes a pipeline to explain RNNs using decision lists over skipgrams, creating synthetic and real-world datasets for evaluation.
result Persistently achieves high explanation fidelity and interpretable rules.
Proposes a method to avoid excessive exploration in reinforcement learning.
problem Avoiding excessive exploration in reinforcement learning to deploy it in practice.
method Designs a novel algorithm using UCB reinforcement learning policy with adaptive exploration constraints.
result Proves that the approach remains conservative while minimizing regret in tabular settings and validates on real-world tasks.
New RL method handles hidden actions in offline learning.
problem Learning from unseen actions in real-world RL datasets.
method LURE (Learning from the Unseen: Robust Estimator) method using next-state variable as proxy.
result Valid statistical inference and improved RL conclusions with hidden actions.
Proposes a method to improve treatment policies in data-scarce clinical settings.
problem Improving treatment policies in data-scarce clinical settings with unobserved confounding.
method Uses a causal mechanism to model the underlying generative process and augments counterfactual trajectories with source domain priors.
result Significantly improves treatment policy performance in a simulated sepsis treatment task.
New approach transfers rewards learned in one environment to reinforcement learning in a new environment.
problem Transfer of rewards learned using inverse reinforcement learning from one environment to a new, different environment.
method Formulate the problem as a joint system of Bellman equations, develop minimax estimators for the target soft-q-function, solve the source and target system of equations jointly. result The coupled approach removes the first-order influence of source Bellman residual error compared to the sequential approach.
Predictive models are finding an increasing number of applications in many industries. As a result, a practical means for trading-off the cost of deploying a model versus its effectiveness is needed. Our work is motivated by risk prediction problems in healthcare. Cost-structures in domains such as healthcare are quite…