Data used for figures of the manuscript, "Demonstration of the optically-driven plasmonic nanomotor designed by deep learning networks", are presented in this folder. You can find the code at https://github.com/mintaechung/Nanomotor_Predictor_Generator. More data information about machine learning can be found in the 'Supplementary_Info\Machine_learning_progress' folder.

The folder "Validationset" provides the correlation between two variables, 'Real' and 'Predicted'. The validationset was not involved during the training but was used to check the validity of the torque predictor. The values in 'Real' indicate optical torques calculated by SIE. The others in 'Predicted' are the output of the torque predictor after training.

The folder "Loss_functions" provides objective loss functions of the generator, discriminator, and the torque predictor as a function of epoch (1 epoch = 550 steps).

The folder "Randomset_and_Generatedset" shows the torque distribution of the initial randomset made by "\Supplementary_Info\Original_random_geometry" and the output of the nanorotor generator after the 3rd iteration.

