Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials
This is an incremental critique and proposal for improving MLIPs to enhance materials simulation and discovery in computational chemistry and materials science.
The paper argues that current universal machine learning interatomic potentials (MLIPs) are insufficient for real-world materials discovery due to overreliance on DFT training data, inability to perform reliable large-scale simulations, and limited understanding of their capabilities, proposing a shift towards more accurate training methods, metrology tools, and efficient models.
Universal Machine Learning Interactomic Potentials (MLIPs) enable accelerated simulations for materials discovery. However, current research efforts fail to impactfully utilize MLIPs due to: 1. Overreliance on Density Functional Theory (DFT) for MLIP training data creation; 2. MLIPs' inability to reliably and accurately perform large-scale molecular dynamics (MD) simulations for diverse materials; 3. Limited understanding of MLIPs' underlying capabilities. To address these shortcomings, we aargue that MLIP research efforts should prioritize: 1. Employing more accurate simulation methods for large-scale MLIP training data creation (e.g. Coupled Cluster Theory) that cover a wide range of materials design spaces; 2. Creating MLIP metrology tools that leverage large-scale benchmarking, visualization, and interpretability analyses to provide a deeper understanding of MLIPs' inner workings; 3. Developing computationally efficient MLIPs to execute MD simulations that accurately model a broad set of materials properties. Together, these interdisciplinary research directions can help further the real-world application of MLIPs to accurately model complex materials at device scale.