TY - GEN
T1 - Comparison of Data-Driven and Physics-Informed Neural Networks for Surrogate Modelling of the Huxley Muscle Model
AU - Milićević, Bogdan
AU - Ivanović, Miloš
AU - Stojanović, Boban
AU - Milošević, Miljan
AU - Simić, Vladimir
AU - Kojić, Miloš
AU - Filipović, Nenad
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.
PY - 2024
Y1 - 2024
N2 - Biophysical muscle models based on sliding filament and cross-bridge theory are called Huxley-type muscle models. The method of characteristics is typically used to solve Huxley’s muscle contraction equation, which describes the distribution of attached myosin heads to the actin-binding sites, called cross-bridges. Once this equation is solved, we can determine the generated force and the stiffness of the muscle fibers, which can then be used at the macro level during finite element analysis. In our paper, we present alternative approaches to finding an approximate solution of Huxley’s muscle contraction equation using neural networks. In one approach, we collect the data from simulations and train multilayer perceptrons to predict probabilities of cross-bridge formation based on the available actin site positions, time, activation, current and previous stretch. In another approach, besides using the data, we also inform the neural network with Huxley’s equation, thus improving the generalization of the neural network’s predictions.
AB - Biophysical muscle models based on sliding filament and cross-bridge theory are called Huxley-type muscle models. The method of characteristics is typically used to solve Huxley’s muscle contraction equation, which describes the distribution of attached myosin heads to the actin-binding sites, called cross-bridges. Once this equation is solved, we can determine the generated force and the stiffness of the muscle fibers, which can then be used at the macro level during finite element analysis. In our paper, we present alternative approaches to finding an approximate solution of Huxley’s muscle contraction equation using neural networks. In one approach, we collect the data from simulations and train multilayer perceptrons to predict probabilities of cross-bridge formation based on the available actin site positions, time, activation, current and previous stretch. In another approach, besides using the data, we also inform the neural network with Huxley’s equation, thus improving the generalization of the neural network’s predictions.
KW - Huxley muscle model
KW - multi-scale modeling
KW - numerical solving of partial differential equations
KW - physics-informed neural networks
UR - http://www.scopus.com/inward/record.url?scp=85195465414&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85195465414&partnerID=8YFLogxK
U2 - 10.1007/978-3-031-60840-7_5
DO - 10.1007/978-3-031-60840-7_5
M3 - Conference contribution
AN - SCOPUS:85195465414
SN - 9783031608391
T3 - Lecture Notes in Networks and Systems
SP - 33
EP - 37
BT - Applied Artificial Intelligence 2
A2 - Filipović, Nenad
PB - Springer Science and Business Media Deutschland GmbH
T2 - 2nd Serbian International Conference on Applied Artificial Intelligence, SICAAI 2023
Y2 - 19 May 2023 through 20 May 2023
ER -