Integrating Small Language Models with ZR+ Digital Coherent Optical Transceiver for Adaptive Signal Quality Optimization
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This paper introduces a framework that integrates Small Language Models (SLMs) into the supervisory control loop of digital coherent optical transceivers. Unlike deep-learning (DL) equalizers or large language models, SLMs offer lightweight, deterministic, on-device inference suitable for embedded control processors in ZR/ZR+ modules. The approach uses SLMdriven reasoning on real-time telemetry to adapt digital signal processor (DSP) parameters, launch power, and impairmentmitigation settings without altering the high-speed symbol-rate pipeline. Experiments show improvements in the Q-factor and a reduction in the pre-forward-error-correction (FEC) bit-error rate (BER) compared to conventional heuristic optimization. These results demonstrate that SLMs enable explainable, lowpower, AI-assisted performance optimization for next-generation coherent optical communication systems.
Fields of Study & Subjects
Optical Network Technologies
Optical Wireless Communication Technologies
Semiconductor Lasers and Optical Devices
Publication details
- Journal / source
- International Conference on Intelligent Systems for Communication, IoT and Security (ICISCoIS)
- Pages
- 1–7
- Publication date
- 8 January 2026
- Language
- EN