[18a-N304-1~8] 3.7 Laser processing

[18a-N304-1] A Deep Learning Approach for Predicting Laser-induced Plasma Characteristics in Glass

〇(M2)Kohei Shimahara¹, Shuntaro Tani¹, Yohei Kobayashi¹ (1.ISSP, Univ. Tokyo)

Keywords:glass, deep learning, plasma

Deep learning is a tool that enables one to approximate functions that A) are highly non-linear, and B) have multi-dimensional inputs such as morphological data frequently acquired in laser processing research; therefore, it has great potential in dealing with complex systems such as laser-matter interaction. Our goal is to predict the time evolution of laser-induced destruction in glass using this technology. In this presentation, I will report our work on designing and training a neural network that can predict the characteristics of laser-induced plasma in glass under a given condition(crater shape and laser power).