Determine the polymer model and its underlying parameters from experimental data

Determine the polymer model and its underlying parameters for an experiment, potentially by learning these quantities as outputs of the neural-network-based reconstruction method.

Background

The proposed trajectory-consistent neural-network method reconstructs the polymer conformation tensor from vorticity measurements without requiring offline conformation data. This feature could permit the model formulation and its parameters to be inferred jointly with the conformation field when the appropriate constitutive description is not known in advance.

The paper identifies the polymer model and its underlying parameters as a key experimental unknown and notes that learning them as network outputs is only a prospective direction being actively explored, rather than a problem resolved in the study.