Improving Robustness and Training Efficiency of Machine-Learned Potentials by Incorporating Short-Range Empirical Potentials

Machine learning force fields (MLFFs) are powerful tools for materials modeling, but their performance is often limited by training data set quality, particularly the lack of rare event configurations. This limitation undermines their accuracy and robustness in long- and large-scale molecular dynamics simulations. In this work, we demonstrate a simple yet general approach to improving robustness and training efficiency by utilizing a hybrid MLFF framework that integrates an empirical short-range repulsive potential. Using solid electrolyte Li₇La₃Zr₂O₁₂ (LLZO) as a model system, we show that purely data-driven MLFFs fail to prevent unphysical atomistic clustering in extended simulations due to inadequate short-range repulsion. In contrast, the hybrid force field eliminates these artifacts, enabling stable long-time simulations, which are critical for studying various properties of LLZO. Our approach also reduces the need for extensive active learning and performs well with just 25 training configurations. By combining physics-driven constraints with data-driven flexibility, this approach is compatible with most existing MLFF architectures and establishes a universal paradigm for developing robust, training-efficient force fields for complex material systems.