Presentation Information

[WP-D-2]An Interpretable Graph Neural Network Approach for Laser Lifetime Prediction

○Yizheng Zuo1,2,3, Ye Zhu1,2,3, Yi Lai1,2,3, Qing Lan1,2,3, Xuwen Liang1,2,3 (1Univ. of Chinese Academy of Sciences, 2Innovation Academy for Microsatellites of CAS, 3Key Laboratory for Satellite Digitalization Technology of CAS)

Keywords:

Semiconductor lasers

This paper proposes an interpretable model based on an attention mechanism and graph convolutional neural network (GCN) for predicting the mean-time-to-failure (MTTF) of lasers, enhancing interpretability while outperforming traditional methods.