Learning dynamics for systems with space structure. Comment on the paper "Collective learning modeling based on the kinetic theory of active particles" by D. Burini et al