#machine learning
Jan 2026
Learning and extrapolating scale-invariant processes
This work explores the problem of prediction in the extrapolation regime of scale-free processes, and considers two paradigmatic problems that are statistically self-similar: the Abelian sandpile model and a two-dimensional fractional Gaussian field obeying linear dynamics.
Anaclara Alvez-Canepa, Cyril Furtlehner, François P. Landes
· Physical Review E · 1 citation