Alert button
Picture for Eric K. Oermann

Eric K. Oermann

Alert button

Language Model Classifier Aligns Better with Physician Word Sensitivity than XGBoost on Readmission Prediction

Nov 15, 2022
Grace Yang, Ming Cao, Lavender Y. Jiang, Xujin C. Liu, Alexander T. M. Cheung, Hannah Weiss, David Kurland, Kyunghyun Cho, Eric K. Oermann

Figure 1 for Language Model Classifier Aligns Better with Physician Word Sensitivity than XGBoost on Readmission Prediction
Figure 2 for Language Model Classifier Aligns Better with Physician Word Sensitivity than XGBoost on Readmission Prediction
Figure 3 for Language Model Classifier Aligns Better with Physician Word Sensitivity than XGBoost on Readmission Prediction
Figure 4 for Language Model Classifier Aligns Better with Physician Word Sensitivity than XGBoost on Readmission Prediction
Viaarxiv icon

Predictive Modeling in the Presence of Nuisance-Induced Spurious Correlations

Jul 06, 2021
Aahlad Puli, Lily H. Zhang, Eric K. Oermann, Rajesh Ranganath

Figure 1 for Predictive Modeling in the Presence of Nuisance-Induced Spurious Correlations
Figure 2 for Predictive Modeling in the Presence of Nuisance-Induced Spurious Correlations
Figure 3 for Predictive Modeling in the Presence of Nuisance-Induced Spurious Correlations
Figure 4 for Predictive Modeling in the Presence of Nuisance-Induced Spurious Correlations
Viaarxiv icon

Confounding variables can degrade generalization performance of radiological deep learning models

Jul 13, 2018
John R. Zech, Marcus A. Badgeley, Manway Liu, Anthony B. Costa, Joseph J. Titano, Eric K. Oermann

Figure 1 for Confounding variables can degrade generalization performance of radiological deep learning models
Figure 2 for Confounding variables can degrade generalization performance of radiological deep learning models
Figure 3 for Confounding variables can degrade generalization performance of radiological deep learning models
Figure 4 for Confounding variables can degrade generalization performance of radiological deep learning models
Viaarxiv icon