Active Meta-Learning for Few-Shot QoT Estimation in Optical Networks
Accurate quality of transmission (QoT) estimation, particularly generalized signal-to-noise ratio (GSNR) prediction, in newly deployed C+L-band optical networks is hindered by data scarcity and incomplete physical link information, leading to a cold-start problem for conventional deep learning methods. To address this...