Fig. 1: Transfer learning of polymer property predictions using all-atom classical MD simulations. | npj Computational Materials

Fig. 1: Transfer learning of polymer property predictions using all-atom classical MD simulations.

From: Scaling law of Sim2Real transfer learning in expanding computational materials databases for real-world predictions

Fig. 1

a Neural network architecture. b Scaling behavior of Sim2Real transfer for four different properties, namely refractive index, density, specific heat capacity (CP), and thermal conductivity. The horizontal axis represents the simulation data size, and the vertical axis shows the MAE averaged over 500 independent trials with 90% confidence interval calculated by performing bootstrapping sampling. The dashed line is the estimated power-law with the estimated equation given at the bottom left, and the horizontal red line indicates the mean MAE for direct learning with no pretraining.

Back to article page