Is Variational Monte Carlo Robust? Sharp Moment Thresholds and Heavy-tailed Stochastic Optimization

arXiv·low signal

Philipp Grohs and Davide Nobile analyze the robustness of Variational Monte Carlo under heavy-tailed noise, deriving sharp moment thresholds for the stochastic optimization used in modern neural-network ansätze in electronic-structure theory. The result gives theoretical convergence guarantees for VMC. A narrow scientific-computing contribution with limited general ML applicability.