In Advance / NanoLabo v.2.5 , the function to control the "self-learning hybrid Monte Carlo method" installed in Advance / NeuralMD Ver.1.6 has been added.
In Advance / NanoLabo v.2.5 , the function to control the "self-learning hybrid Monte Carlo method" installed in Advance / NeuralMD Ver.1.6 has been added.
The self-learning hybrid Monte Carlo method is a first-principles Monte Carlo algorithm 1) developed by the Japan Atomic Energy Agency . By applying the trajectory of molecular dynamics calculation by Neural Network force field as the proposed structure in the Monte Carlo method, it is possible to realize efficient structural sampling while guaranteeing the accuracy of the first principle for the Monte Carlo calculation itself. Simultaneously with the execution of the Monte Carlo calculation, the learning of the Neural Network force field is also performed in parallel using the result of the first-principles calculation calculated for each structure. As a result, when the method is executed, a Neural Network force field specialized for the target system will be automatically generated.
1) Y. Nagai, et al., Rhys. Rev. B 102 041124 (2020)
Benefits of self-learning hybrid Monte Carlo method
By linking with the environment of Mat3ra or Azure Cycle Cloud , you can utilize the self-learning hybrid Monte Carlo method more effectively. The environment has already been set for Mat3ra and Azure Cycle Cloud.
Original Source from: https://ctc-mi-solution.com/advance-nanolabo-v-2-5-リリース/
