• A Hybrid Conference – In-Person and Virtual Presentations
  • Technical Conference:  24 – 28 March 2024
  • Exhibition: 26 – 28 March 2024
  • San Diego Convention Center, San Diego, California, USA

Neural Networks for Optical Fiber Transmission: Hype or Hope?

Sunday, 24 March, 13:00 – 15:30

Room 7

Neural network (NN) based machine learning has been investigated for applications in optical fiber communications for many years. The studies covered almost all aspects of the field, including optical transmission modeling, optical link optimization, linear and nonlinear impairment mitigation, etc. Despite the extensive research, practical applications of neural network-based machine learning in optical fiber communications remain elusive. This workshop would like to discuss a crucial question: Can NNs replicate their successes in this field as they have in natural language processing with ChatGPT?

The workshop shall cover but not be limited to the following scopes:

1. Implementation complexity challenges faced by the NN-based machine learning to be practical for high-speed optical communications
2. Requirements on order and length of the PRBS dataset employed for the training and testing;
3. Requirement on training and testing dataset ratio
4. Verification of the tracking bandwidth of the NN for dynamic transmission system effects
5. Ethical considerations of using NN-based machine learning in optical fiber communications, such as data privacy and security concerns

Organizers

Ming-Fang (Yvonne) Huang, NEC Labs America, United States

Yi Cai, Soochow University, China

Amirhossein Ghazisaeidi, Nokia Bell Labs, France

Speakers

Alan Lau, Hong Kong Polytechnic University, Hong Kong

Takeo Sasai, NTT Network Innovation Labs, Japan

Faith Yaman, NEC Labs America, United States

Lilin Yi, Shanghai Jiao Tong University, China