Welcome to [Daby-Seesaram, Škardová, Genet, Submitted]’s demos!

• These demos are based on the NeuROM-py code version 3.1.17 and allow reproducing most figures of our paper [Daby-Seesaram, Škardová, Genet, Submitted].

• The notebooks can be executed online with binder (no download required).

Hybridising standard reduced-order modelling methods with interpretable sparse neural networks

This work proposes hybridising classical model-order reduction methods with machine learning capabilities to provide real-time solutions to mechanics problem.

Analogous to techniques like the Proper Generalised Decomposition (PGD) or the Higher Order Singular Value Decomposition (HOSVD), the parametric mechanical field is represented through a tensor decomposition, effectively mitigating the curse of dimensionality associated with numerous parameters. Each mode of the tensor decomposition is given by the output of a sparse neural network within the HiDeNN framework, constraining the weights and biases to emulate classical shape functions used in the Finite Element Method.