Tag: MIMIC-III

A new open-access study reports that fusing attribute features with graph structure and training classifiers collaboratively reached 0.724 accuracy on ...

A new CARE-Med framework combining dense retrieval, natural language inference verification, and confidence scoring cut unsupported-response rates in medical LLMs ...

Researchers have developed a deep reinforcement learning framework that generates synthetic patient data and explicitly avoids high-risk clinical states, outperforming ...

A new study transferred ICU mortality prediction models between two clinical databases and found that fixed decision thresholds can silently ...

A new hybrid deep learning architecture combining a fine-tuned U-Net with a time-weighted multi-task bi-directional LSTM significantly improves real-time detection ...

Researchers built an eight-variable machine learning model that predicts in-hospital mortality in patients with sepsis-related myocardial injury and validated its ...

An exploratory MIMIC-III cohort study found that baseline thrombocytopenia showed no precise independent association with a 48-hour respiratory-support burden endpoint ...

Researchers in China have developed MKMed, a cross-modal AI framework that aligns five types of drug knowledge to overcome the ...