⚕️🚑 1 Paper accepted to IEEE TBME 2025! 🚑⚕️
We are thrilled to share that our lab has a paper accepted to IEEE Transactions on Biomedical Engineering (TBME) 2025! A big congratulations to our joint first authors, Ahmet Demirkaya and Kyle Lockwood, for their outstanding work:
- ⭐
A Hybrid ODE-NN Framework for Modeling Incomplete Physiological Systems
A Hybrid ODE-NN Framework for Modeling Incomplete Physiological Systems
A Hybrid ODE-NN Framework for Modeling Incomplete Physiological Systems
Ahmet and Kyle proposed a hybrid modeling framework that augments partially known physiological Ordinary Differential Equation (ODE) systems with neural networks. Using recursive Bayesian estimation, this approach tracks unmeasured physiological states and approximates the outputs of missing ODEs, even under noisy conditions. The framework was validated on neuron and retinal circulation models, demonstrating improved accuracy in modeling physiological systems. By preserving biophysical principles and leveraging system dynamics, it provided interpretable insights into the underlying processes, offering advantages over purely data-driven models.