抄録
Machine learning, one of the methods in artificial intelligence, is computer algorithm for automatically
improving by using the experience and/or the data. Recently, the study on the application of this algorithm
is actively occurred in both CW and pulsed laser processing. This report provides an example of
applying this algorithm to predict the ablation efficiency of femtosecond laser processing. For the work,
five materials are selected: cold-rolled steel sheet and aluminum as the metal, Silicon wafer as the semiconductor,
glass as the insulator, and polycrystalline diamond as the composite material. As the results,
the ablation efficiencies of all materials were predicted with the accuracies of up to ±40%. In addition,
the two application challenges of the learning model are introduced: (1) prediction of the ablation efficiency
of one material using machine learning models from material properties and the other material
processing results, and (2) influence of manufacturing errors on laser processing results.
