Long Short Term Memory Regression and Dense Neural Network Enabled 200-Gbit/s PAM4 using GeSi Electro-Absorption Modulator (EAM)
A long-short-term-memory-regression (LSTM) and dense-neural-network (DNN) enabled high-speed four-level pulse-amplitude-modulation (PAM4) generation using GeSi electro-absorption modulator (EAM) is demonstrated. It supports 200-Gbit/s 1-km standard-single-mode-fiber (SSMF) and 240-Gbit/s back-to-back (B2B) transmissions fulfilling the soft-decision-forward-error-correction (SD-FEC).
Authors
- Hon Ki Tsang (ORCID: https://orcid.org/0000-0003-2777-1537)
- D. Chan (ORCID: https://orcid.org/0000-0002-3526-5389)
- Chi‐Wai Chow (ORCID: https://orcid.org/0000-0002-7488-2611)
- Yuan-Zeng Lin (ORCID: https://orcid.org/0000-0002-1290-5419)
- Weilin Chen (ORCID: https://orcid.org/0000-0002-6259-5834)
- Tun-Yao Hung
Institutions
- National Yang Ming Chiao Tung University (TW)
- Chinese University of Hong Kong (HK)
Publication Details
- Published
- 2025-04-14
- DOI
- https://doi.org/10.1109/siphotonics64386.2025.10985562
- Primary Topic
- Photonic and Optical Devices
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- Innovation and Technology Fund
- National Science and Technology Council