Spatially resolved macroscopic near-threshold harmonics generated from harmoniclike and resonant transitions using an artificial neural network

We investigate the generation of near-threshold harmonics (NTHs) based on solving the three-dimensional time-dependent Schr\"odinger equation and utilizing an artificial neural network as a surrogate model to study the macroscopic propagation. Through our research, we identify two distinct pathways for NTH generation: harmoniclike transition and resonant transition. These pathways exhibit varying laser-parameter dependencies and phase-matching conditions, making it possible to spatially separate and selectively enhance NTHs by manipulating laser parameters in experimental setups. Our study provides insights into the underlying physics of the NTH macroscopic propagation and presents a methodology that can be applied to computationally intensive problems in the field of strong-field physics.