The data presented in the paper "Mechanical Property Prediction of Random Copolymers Using Uncertainty-based Active Learning," pertains to the utilization of active learning to construct a surrogate model for random copolymers. Initially, the model was trained using data from reference [1], which consisted of block copolymer data. Subsequently, we employed active learning to sample random copolymers and labeled them. Following this, we refined the model through iterative teaching, resulting in significant improvements. The data available for sampling is provided in this repository.
All the data is saved as a .npy file. The first array contains the seven input parameters, while the second array contains the stress values corresponding to strain from 0 to 3.
[1] Aoyagi, T. (2022). Optimization of the elastic properties of block copolymers using coarse-grained simulation and an artificial neural network. Computational Materials Science, 207, 111286.