Simulation, Modeling and Prediction of a Pharmacodynamic Animal Tissue Culture Compartment Model by Physical Informed Neural Network

TL;DR

This paper demonstrates that Physical Informed Neural Networks (PINNs) can effectively model, simulate, and predict pharmacodynamic cell culture compartment models, outperforming traditional methods especially with large datasets.

Contribution

The work introduces a PINN-based approach using PyTorch to model and predict cell culture compartment models, showing improved accuracy and potential for handling large datasets.

Findings

Achieved a low loss value of 0.0004853 in model predictions.

PINNs can serve as effective tools for complex biological compartment models.

Potential for better performance with large datasets.

Abstract

Compartment models of cell culture are widely used in cytology, pharmacology, toxicology and other fields. Numerical simulation, data modeling and prediction of compartment models can be realized by traditional differential equation modeling methods. At the same time, with the development of software and hardware, Physical Informed Neural Network (PINN) is widely used to solve differential equation models. This work models, simulates and predicts the cell culture compartment model based on the machine learning framework PyTorch with an 16 hidden layers neural network, including 8 linear layers and 8 feedback active layers. The results showed a loss value of 0.0004853 for three-component four-parameter quantitative pharmacodynamic model predictions in this way, which is evaluated by Mean Square Error (MSE). In summary, Physical Informed Neural Network can serve as an effective tool to deal with cell culture compartment models and may perform better in dealing with big datasets.