AI RESEARCH

Generalization of Long-Range Machine Learning Potentials in Complex Chemical Spaces

arXiv CS.LG

ArXi:2512.10989v2 Announce Type: replace-cross The vastness of chemical space makes generalization a central challenge in the development of machine learning interatomic potentials (MLIPs). While MLIPs could enable large-scale atomistic simulations with near-quantum accuracy, their usefulness is often limited by poor transferability to out-of-distribution samples.