Proposal Description

Intensive functional, time series, imaging data, and high-throughput data are widely collected from multiple subjects with high sampling frequencies in nowadays biomedical studies. While machine learning models are excellent in utilizing such data to generate classifications and predictions, scientific interpretations have always been a difficult problem. How to process these signals by developing and applying rigorous techniques from both statistical and engineering methodologies to gain scientific insights into biomedical problems raises both challenges and opportunities. Four experts in statistical and engineering signal processing will present their recent works in this proposed invited paper session.
- Dr. Wensheng Guo from University of Pennsylvania, USA, will present how to extract localized functional features from ECG data.
- Dr. Kun Huang from Indiana University School of Medicine, USA, will present how to process histopathology images and integrate them with genomic data in precision medicine.
- Dr. Ziyue Liu from Indian University School of Medicine, USA, will present a state space to nonparametric regressions when both the independent and the dependent variables are from multivariate time series data.
- Dr. Ping Ma from University of Georgia, USA, will present a diffusion-process based method to denoise sequence-based spatial transcriptomics.