Overview of the implemented simulation tool and training framework in Diffrax.

<p>Latin Hypercube sampling is used to initialize parameters given a lower- and upper bound value (1). The initial conditions are retrieved from the observed dataset (2) and the ODEs are set up based on an imported SBML model, or a self-implemented model (3). After predicting a time-series dat...

Full description

Saved in:

Bibliographic Details
Main Author:	Paul van Lent (16876977) (author)
Other Authors:	Olga Bunkova (21673486) (author), Bálint Magyar (17903285) (author), Léon Planken (21673489) (author), Joep Schmitz (14822563) (author), Thomas Abeel (206461) (author)
Published:	2025
Subjects:	Biophysics Molecular Biology Biotechnology Ecology Biological Sciences not elsewhere classified Information Systems not elsewhere classified ordinary differential equations motivation :</ b many methodological developments model biological systems results :</ b hybridizing kinetic model metabolic kinetic models hybridizing kinetic models implementation :</ b models using jax time compilation capabilities training framework implementation kinetic models complex model training framework robust capabilities modeling time jax allows sbml models widespread use strong platform results show remains challenging recent work reaction mechanism python package prior information parameter initialization neural odes neural networks neural network large collection furthermore showcase different degrees automatic differentiation also allowing
Tags:	Add Tag No Tags, Be the first to tag this record!

Overview of the implemented simulation tool and training framework in Diffrax.

Similar Items