Cheol Woo Park1,Babak Sadigh1,Yu-Ting Hsu1,Fei Zhou1
Lawrence Livermore National Lab1
Cheol Woo Park1,Babak Sadigh1,Yu-Ting Hsu1,Fei Zhou1
Lawrence Livermore National Lab1
Simulation of microstructure evolution have largely relied on coarse-grained models that obey evolution principles described by partial differential equations. However, quantitatively capturing the precipitation kinetics in these models remains a challenge. In this talk, we show that stochastic dynamics that occur on the coarse-grained level in microstructure evolution can be learned in a data-driven manner using a diffusion probabilistic model (DPM) which enables the model to accurately reenact nucleation and growth. We demonstrate the capability of DPM by replicating the precipitation dynamics of a symmetric Lennard-Jones binary mixture. This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.