| Abstract: |
The REU Site: Deep Learning for Dynamical Systems in Biological and Physical Sciences at the University of Southern Mississippi (USM) will engage ten undergraduate students annually over three summers in an eight-week program. The program focuses on applying deep learning to model complex dynamical systems in epidemiology, biophysics, and astrophysics. Participants will undergo a Python bootcamp and hands-on workshops on neural network architectures (feedforward, convolutional, recurrent, and transformers) and mathematical modeling. They will design neural networks, solve nonlinear ODE systems, collect data, and train models to address real-world problems. The program emphasizes mentorship, teamwork, and professional development, culminating in conference-style presentations. ACCESS resources will support computationally intensive deep learning tasks, enabling students to train models and analyze data efficiently, given USM's limited GPU availability. |
0000-0002-0344-5991