weishengtang.com

Weisheng Tang

EN · 中文

Academic Homepage

Weisheng Tang

Postdoctoral Research Associate, Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxville

Weisheng Tang is a Postdoctoral Research Associate in the Department of Electrical Engineering and Computer Science at the University of Tennessee, Knoxville.

His research interests include machine learning, computer vision, remote sensing, training dynamics, and multimodal biomedical data analysis. His recent work spans predictive training methods for deep neural networks, satellite image time-series analysis, and data-driven methods for scientific and biomedical applications.

About · Publications · Projects · Demos

Research Themes

  • Machine learning and training dynamics
  • Computer vision and remote sensing
  • Multimodal biomedical data analysis

Knoxville, Tennessee, USA

Selected Publications

Enhanced Phase Transition Detection in Satellite Imagery Time Series via a Koopman-Based Framework

Weisheng Tang, F. Wang, M. Tang, K. Georgiou, H. Qi, C. Champion, M. Bosch

Submitted to IEEE Transactions on Geoscience and Remote Sensing · 2025

remote sensing · satellite imagery

Predictive Differential Training Guided by Training Dynamics

F. Wang, Weisheng Tang, L. Harris, H. Qi, W. Dan, I. Mezic

ICLR · 2025

training dynamics · deep learning

Selected Projects

Predictive Training Guided by Training Dynamics

Predictive Training Guided by Training Dynamics

active

Research on faster and more controllable neural network training using training dynamics and Koopman-inspired ideas.

machine learning · training dynamics

Satellite Image Time-Series Analysis

Satellite Image Time-Series Analysis

active

Remote sensing research on construction phase detection and transition analysis from satellite imagery time series.

remote sensing · computer vision

CV

A concise summary of education, positions, and selected experience.

View CV

Demos

The site keeps demos in a separate top-level section so they can grow independently.

1 demo available. Browse demos