FURI | Spring 2020

Neural ODE Modeling for Many-Body Interactions

Data icon, disabled. Four grey bars arranged like a vertical bar chart.

Neural Ordinary Differential Equations (ODEs) present a new pathway for the reversible modeling of many-body systems, including those for the predictive analysis of particle interactions. Such modeling can determine destination trajectories for cancer cell coupling interactions in the extracellular matrix given respective start or end-points, assisting in targeting potential cell cluster convergence for preventative removal. This project is the basis for robust predictive modeling of many-body systems, and a means to optimize ODE fitting for equivariant and stochastic particle systems.

Student researcher

Sreeharsha Lakamsani

Computer science

Hometown: Chandler, Arizona, United States

Graduation date: Spring 2023