Comparing classical ordinary differential equation and neural network models for reduced-order single-cell electrophysiology

dc.contributor.advisor-co1Rocha, Bernardo Martins
dc.contributor.advisor-co1Latteshttp://lattes.cnpq.brpt_BR
dc.contributor.advisor1Santos, Rodrigo Weber dos
dc.contributor.advisor1Latteshttp://lattes.cnpq.brpt_BR
dc.contributor.referee1Lobosco, Marcelo
dc.contributor.referee1Latteshttp://lattes.cnpq.brpt_BR
dc.contributor.referee2Cherry, Elizabeth Maura
dc.contributor.referee2Latteshttp://lattes.cnpq.brpt_BR
dc.creatorWerneck, Yan Barbosa
dc.creator.Latteshttp://buscatextual.cnpq.brpt_BR
dc.date.accessioned2025-01-31T15:28:09Z
dc.date.available2025-01-31
dc.date.available2025-01-31T15:28:09Z
dc.date.issued2024-11-13
dc.description.abstractModeling cardiac electrophysiology plays a crucial role in advancing non-invasive diagnostics and enhancing our understanding of heart function. Historically, models describing excitable cells through systems of Ordinary Differential Equations (ODEs) have been the standard in electrophysiology modeling. These models range from detailed representations of ion channel dynamics to simplified reduced-order models that capture the behavior of excitability phenomenologically. In this work, we compare a fast reduced-order model with data-driven and physics-informed neural networks to assess their effectiveness as efficient replacements for numerical solutions. For this, the FitzHugh-Nagumo model was used, and scenarios with increasing complexity were studied. The networks were trained using numerical data and knowledge of model physics, derived from the ODEs. Additionally, several techniques were employed to improve training, including architecture optimization, increased point density in regions of high error, and time-domain splitting. Inference was conducted using the state-of-the-art TensorRT SDK to speed up model inference, leveraging tensor core matrix-matrix specialization to ensure maximum performance. We observed up to a 1.8x speedup compered to numerical models optimized and implemented in CUDA, with minimal loss in accuracy. These gains highlight valuable use cases for neural network emulators, as faster substitute for numerical methods when complexity can be controlled, while still emphasizing the prominence of equation-based modeling in cardiac electrophysiology in general due to their flexibility.pt_BR
dc.description.resumo-pt_BR
dc.identifier.urihttps://repositorio.ufjf.br/handle/ufjf/18126
dc.languageporpt_BR
dc.publisherUniversidade Federal de Juiz de Fora (UFJF)pt_BR
dc.publisher.countryBrasilpt_BR
dc.publisher.departmentICE – Instituto de Ciências Exataspt_BR
dc.publisher.initialsUFJFpt_BR
dc.publisher.programPrograma de Pós-graduação em Modelagem Computacionalpt_BR
dc.rightsAcesso Abertopt_BR
dc.rightsAttribution 3.0 Brazil*
dc.rights.urihttp://creativecommons.org/licenses/by/3.0/br/*
dc.subjectRedes neuraispt_BR
dc.subjectEDOpt_BR
dc.subjectPotencial de açãopt_BR
dc.subjectPINNspt_BR
dc.subjectModelos surrogadospt_BR
dc.subject.cnpqCNPQ::CIENCIAS EXATAS E DA TERRApt_BR
dc.titleComparing classical ordinary differential equation and neural network models for reduced-order single-cell electrophysiologypt_BR
dc.typeDissertaçãopt_BR

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
yanbarbosawerneck.pdf
Size:
104.68 MB
Format:
Adobe Portable Document Format

License bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description: