[TuG1] Optical Network Optimization

10:00 AM - 10:15 AM JST(1:00 AM - 1:15 AM UTC)

[TuG1-3] Network for AI: Efficient Mapping with Fine Grain OTN for AI Computing Services Via Entropy-Coupled Incremental Learning

○Tiankuo Yu¹, Hui Yang¹, Qiuyan Yao¹, Yang Zhao², Shengye Gong¹, Zepeng Zhang¹, Jie Zhang¹, Mohamed Cheriet³ (1. Beijing University of Posts and Telecommunications (China), 2. China Mobile Research Institute (China), 3. University of Quebec (Canada))

Keywords:

Optical cross-connects/add-drop multiplexers and switching subsystems Grid/cloud computing over optical networks,Artificial intelligence and machine learning for optical network design,control,and management,Optical core/metro network architecture,design,control,and management

We propose an entropy-coupled incremental learning (ECIL) mapping framework for fine grain optical transport network (fgOTN), achieving significant improvements in convergence, latency, and resource utilization, thereby optimizing AI computing service integration and processing efficiency.