Generalized few-shot transfer learning architecture for modeling the EDFA gain spectrum

2025-07-29

AI Summary

Modeling the gain spectra of EDFAs in multi-vendor optical networks relies heavily on extensive empirical measurements and suffers from poor generalizability. Method: This paper proposes a few-shot transfer learning framework featuring a universal transfer architecture that aligns feature spaces across heterogeneous EDFAs via covariance matching loss; it employs a semi-supervised self-normalizing neural network, combining noise-augmented unsupervised pretraining with weighted MSE-based supervised fine-tuning. Contribution/Results: To the best of our knowledge, this is the first work to integrate covariance matching into cross-type EDFA gain prediction, enabling efficient adaptive modeling for both homogeneous and heterogeneous amplifiers. Evaluations on the COSMOS and Open Ireland testbeds demonstrate a substantial reduction in required measurements, a 21.3% decrease in mean absolute error, tighter error distribution, and concurrent improvements in generalization capability and practical applicability.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Accurate modeling of the gain spectrum in Erbium-Doped Fiber Amplifiers (EDFAs) is essential for optimizing optical network performance, particularly as networks evolve toward multi-vendor solutions. In this work, we propose a generalized few-shot transfer learning architecture based on a Semi-Supervised Self-Normalizing Neural Network (SS-NN) that leverages internal EDFA features - such as VOA input or output power and attenuation, to improve gain spectrum prediction. Our SS-NN model employs a two-phase training strategy comprising unsupervised pre-training with noise-augmented measurements and supervised fine-tuning with a custom weighted MSE loss. Furthermore, we extend the framework with transfer learning (TL) techniques that enable both homogeneous (same-feature space) and heterogeneous (different-feature sets) model adaptation across booster, preamplifier, and ILA EDFAs. To address feature mismatches in heterogeneous TL, we incorporate a covariance matching loss to align second-order feature statistics between source and target domains. Extensive experiments conducted across 26 EDFAs in the COSMOS and Open Ireland testbeds demonstrate that the proposed approach significantly reduces the number of measurements requirements on the system while achieving lower mean absolute errors and improved error distributions compared to benchmark methods.

Problem

• Accurate modeling of EDFA gain spectrum for optical networks
• Generalized few-shot transfer learning for EDFA gain prediction
• Addressing feature mismatches in heterogeneous transfer learning scenarios

Innovation

• Semi-Supervised Self-Normalizing Neural Network
• Two-phase training with noise-augmented data
• Transfer learning with covariance matching loss