Erina Paul

This paper introduces Coracle, a conformalized framework for multimodal AI that provides valid uncertainty estimates, outperforming existing methods in uncertainty estimation and calibration.

This paper introduces BayesCOOP, a Bayesian method for multimodal integration that avoids MCMC and variational approaches, showing improved performance in estimation and uncertainty assessment.

The paper introduces an ensemble framework for differential analysis that combines multiple models to improve performance across various omics data types.

The paper introduces a Bayesian ensemble method for multi-omics prediction and classification, showing improved performance and uncertainty quantification across four datasets.