Machine-learning interatomic potentials enable first-principles multiscale modeling of lattice thermal conductivity in graphene/borophene heterostructures

We highlight that machine-learning interatomic potentials trained over short AIMD trajectories enable first-principles multiscale modeling, bridging DFT level accuracy to the continuum level and empowering the study of complex/novel nanostructures.