• Technical Conference:  30 March – 03 April 2025
  • Exhibition: 01 – 03 April 2025
  • Moscone Center, San Francisco, California, USA
Home > Poster / Demo

Poster / Demo

Data Labeling Using Unsupervised Cascaded Pre-Training With Fused Multi-Port Data for Optical Failure Management (W2B.8)

Presenter: Weijie Yang, Beijing University of Posts and Telecommunications

We propose an unsupervised cascaded pre-training data labeling method that considers the intrinsic correlation of multi-port data, and verifies the scheme validity in failure prediction using real multi-port data from optical networks.

Authors:Weijie Yang, Beijing University of Posts and Telecommunications / Chunyu Zhang, Beijing University of Posts and Telecommunications / Danshi Wang, Beijing University of Posts and Telecommunications / Zhu Hong, The Intelligent Network Innovation Center of Chinaunicom / Xinxing Xu, The Intelligent Network Innovation Center of Chinaunicom / Degang Shi, China United Network Communications Group Co., Ltd. / Min Zhang, Beijing University of Posts and Telecommunications


Poster Presentation

Authentication Required

Access to poster content is available to logged in registered attendees, please login now to participate.