New study design elicits unobserved confounders by querying experts on treatment intent for matched units, with theoretical conditions and a proof-of-concept demonstration in ICU electronic health records using clinical notes as proxy.
hub
MIMIC-III, a freely accessible critical care database
21 Pith papers cite this work, alongside 8,137 external citations. Polarity classification is still indexing.
21
Pith papers citing it
8,137
external citations · OpenAlex
hub tools
citation-role summary
background 1
method 1
citation-polarity summary
representative citing papers
CheckMIABench converts LLMs with intermediate checkpoints into clean MIA testbeds by using pre- and post-checkpoint training data from the same distribution and evaluates published attacks on Pythia and OLMo models while releasing an open-source library.
CopFITi is the first marginalization-consistent copula for irregular multivariate time series, using normalizing flows for marginals and a Gaussian mixture copula for dependencies to reach new state-of-the-art joint density modeling.
Presents first online L2D algorithm for multiclass classification with bandit feedback and varying experts, achieving O((n+n_e)T^{2/3}) regret generally and O((n+n_e)√T) under low noise.
Physiome-ODE is a new benchmark consisting of 50 IMTS datasets derived from biological ODEs that shows ODE-based forecasting models performing better and differentiating more meaningfully than on the existing four datasets.
eCREAM-MedCorpus releases ~4M anonymized Italian ED clinical notes and a 6k-note 132-item CRF annotation set, with zero-shot Gemma/MedGemma CRF-filling baselines.
OpenRFM combines a relational transformer backbone with a batch-level ICL layer and homophily-aware synthetic-plus-real pre-training to improve relational in-context learning by ~30% over prior open models and surpass KumoRFMv1.
MedGym introduces a continuous-time RL benchmark for medical treatment derived from clinical data via PINNs, supporting offline/online evaluation on personalization, safety, and discrete vs continuous methods.
DeconDTN-Toolkit simulates provenance shifts to expose ERM vulnerabilities and provides tools plus a robust OOD indicator for mitigating confounding by data provenance.
A cross-modal masked autoencoder creates reusable biosignal fingerprints that match or exceed specialist models on seven cardiovascular tasks using only single-modality input.
Counterfactual prompting effects on LLMs are often indistinguishable from those caused by meaning-preserving paraphrases, causing most previously reported demographic sensitivities to disappear under proper statistical comparison.
FeatEHR-LLM uses LLMs with tool-augmented code generation on dataset schemas to extract clinically meaningful features from irregular EHR time series, achieving the highest AUROC on 7 of 8 ICU prediction tasks with gains up to 6 points over baselines.
ConStruM improves LLM-based schema matching by using a context tree and global similarity hypergraph to assemble query-specific evidence packs from available schema metadata.
TabICL scales in-context learning to large tabular data via column-then-row attention for row embeddings followed by a transformer, matching TabPFNv2 speed and performance while outperforming it and CatBoost on datasets over 10K samples.
Introduces Interaction-Limited Safe Continuous-Time RL reformulating medical treatment as an option-based SMDP with trajectory-level safety guarantees and finite-sample learning bounds.
Single-agent LLM frameworks outperform naive multi-agent systems in multimodal clinical risk prediction tasks and are better calibrated.
LLMs match or beat supervised BERT models on detecting whether a discharge note contains an actionable clinical task but trail on classifying the exact type of action, pointing to the need for datasets that explain why each span was labeled actionable.
SurvBench supplies a configurable, open-source preprocessing pipeline that standardizes multi-modal EHR data from four critical-care databases for single-risk and competing-risk survival analysis.
A review synthesizes evidence from EEG, EMG, ECG, PPG and ocular signals to argue that waveform morphology, rather than modality or model class, primarily determines TSC performance and interpretability.
The paper introduces ClinQueryAgent, a conversational agent that converts natural language queries into database queries for population health management while keeping patient data secure, and reports its use by 128 staff across 15 NHS practices covering 148,319 patients.
SNOMED-CT graph embeddings via random walks and Poincaré methods yield 5-6x better concept similarity and 6-20% better patient diagnosis prediction than prior embeddings.
citing papers explorer
-
Confounder Detection via Treatment Intent: A New Observational Study Design
New study design elicits unobserved confounders by querying experts on treatment intent for matched units, with theoretical conditions and a proof-of-concept demonstration in ICU electronic health records using clinical notes as proxy.
-
CheckMIABench: Firm Foundations For Membership Inference Attacks on Language Models
CheckMIABench converts LLMs with intermediate checkpoints into clean MIA testbeds by using pre- and post-checkpoint training data from the same distribution and evaluates published attacks on Pythia and OLMo models while releasing an open-source library.
-
Valid and Expressive Copulas for Irregular Multivariate Time Series
CopFITi is the first marginalization-consistent copula for irregular multivariate time series, using normalizing flows for marginals and a Gaussian mixture copula for dependencies to reach new state-of-the-art joint density modeling.
-
Online Learning-to-Defer with Varying Experts
Presents first online L2D algorithm for multiclass classification with bandit feedback and varying experts, achieving O((n+n_e)T^{2/3}) regret generally and O((n+n_e)√T) under low noise.
-
Physiome-ODE: A Benchmark for Irregularly Sampled Multivariate Time Series Forecasting Based on Biological ODEs
Physiome-ODE is a new benchmark consisting of 50 IMTS datasets derived from biological ODEs that shows ODE-based forecasting models performing better and differentiating more meaningfully than on the existing four datasets.
-
eCREAM-MedCorpus A Large-Scale Corpus of Clinical Notes for Italian
eCREAM-MedCorpus releases ~4M anonymized Italian ED clinical notes and a 6k-note 132-item CRF annotation set, with zero-shot Gemma/MedGemma CRF-filling baselines.
-
OpenRFM: Dissecting Relational In-Context Learning
OpenRFM combines a relational transformer backbone with a batch-level ICL layer and homophily-aware synthetic-plus-real pre-training to improve relational in-context learning by ~30% over prior open models and surpass KumoRFMv1.
-
MedGym:A Unified Continuous-Time Benchmark for Dynamic Medical Treatment Reinforcement Learning
MedGym introduces a continuous-time RL benchmark for medical treatment derived from clinical data via PINNs, supporting offline/online evaluation on personalization, safety, and discrete vs continuous methods.
-
DeconDTN-Toolkit: A Library for Evaluation and Enhancement of Robustness to Provenance Shift
DeconDTN-Toolkit simulates provenance shifts to expose ERM vulnerabilities and provides tools plus a robust OOD indicator for mitigating confounding by data provenance.
-
Biosignal Fingerprinting: A Cross-Modal PPG-ECG Foundation Model
A cross-modal masked autoencoder creates reusable biosignal fingerprints that match or exceed specialist models on seven cardiovascular tasks using only single-modality input.
-
Compared to What? Baselines and Metrics for Counterfactual Prompting
Counterfactual prompting effects on LLMs are often indistinguishable from those caused by meaning-preserving paraphrases, causing most previously reported demographic sensitivities to disappear under proper statistical comparison.
-
FeatEHR-LLM: Leveraging Large Language Models for Feature Engineering in Electronic Health Records
FeatEHR-LLM uses LLMs with tool-augmented code generation on dataset schemas to extract clinically meaningful features from irregular EHR time series, achieving the highest AUROC on 7 of 8 ICU prediction tasks with gains up to 6 points over baselines.
-
ConStruM: A Structure-Guided LLM Framework for Context-Aware Schema Matching
ConStruM improves LLM-based schema matching by using a context tree and global similarity hypergraph to assemble query-specific evidence packs from available schema metadata.
-
TabICL: A Tabular Foundation Model for In-Context Learning on Large Data
TabICL scales in-context learning to large tabular data via column-then-row attention for row embeddings followed by a transformer, matching TabPFNv2 speed and performance while outperforming it and CatBoost on datasets over 10K samples.
-
Interaction-Limited Safe Continuous-Time RL for Dynamical Medical Treatment
Introduces Interaction-Limited Safe Continuous-Time RL reformulating medical treatment as an option-based SMDP with trajectory-level safety guarantees and finite-sample learning bounds.
-
AgentRx: A Benchmark Study of LLM Agents for Multimodal Clinical Prediction Tasks
Single-agent LLM frameworks outperform naive multi-agent systems in multimodal clinical risk prediction tasks and are better calibrated.
-
Systematic Evaluation of Large Language Models for Post-Discharge Clinical Action Extraction
LLMs match or beat supervised BERT models on detecting whether a discharge note contains an actionable clinical task but trail on classifying the exact type of action, pointing to the need for datasets that explain why each span was labeled actionable.
-
SurvBench: A Standardised Preprocessing Pipeline for Multi-Modal Electronic Health Record Survival Analysis
SurvBench supplies a configurable, open-source preprocessing pipeline that standardizes multi-modal EHR data from four critical-care databases for single-risk and competing-risk survival analysis.
-
Modality vs. Morphology: A Framework for Time Series Classification for Biological Signals
A review synthesizes evidence from EEG, EMG, ECG, PPG and ocular signals to argue that waveform morphology, rather than modality or model class, primarily determines TSC performance and interpretability.
-
ClinQueryAgent: A Conversational Agent for Population Health Management
The paper introduces ClinQueryAgent, a conversational agent that converts natural language queries into database queries for population health management while keeping patient data secure, and reports its use by 128 staff across 15 NHS practices covering 148,319 patients.
-
Snomed2Vec: Random Walk and Poincar\'e Embeddings of a Clinical Knowledge Base for Healthcare Analytics
SNOMED-CT graph embeddings via random walks and Poincaré methods yield 5-6x better concept similarity and 6-20% better patient diagnosis prediction than prior embeddings.