[Paper Review] Semantic optical fiber communication system

Zhenming Yu, Hongyu Huang+4 more|arXiv (Cornell University)|Dec 27, 2022

Neural Networks and Reservoir Computing4 citations

TL;DR

This paper proposes a semantic optical fiber communication (SOFC) system that replaces traditional bit-based transmission with deep learning–derived semantic symbols, achieving higher spectral efficiency and improved robustness under low optical power and impairments. Experimental results show superior compression and stability compared to conventional systems.

ABSTRACT

The current optical communication systems minimize bit or symbol errors without considering the semantic meaning behind digital bits, thus transmitting a lot of unnecessary information. We propose and experimentally demonstrate a semantic optical fiber communication (SOFC) system. Instead of encoding information into bits for transmission, semantic information is extracted from the source using deep learning. The generated semantic symbols are then directly transmitted through an optical fiber. Compared with the bit-based structure, the SOFC system achieved higher information compression and a more stable performance, especially in the low received optical power regime, and enhanced the robustness against optical link impairments. This work introduces an intelligent optical communication system at the human analytical thinking level, which is a significant step toward a breakthrough in the current optical communication architecture.

Motivation & Objective

• To address the inefficiency of traditional optical communication systems that transmit redundant bits without considering semantic meaning.
• To reduce information redundancy by extracting and transmitting only essential semantic content using deep learning.
• To improve system robustness under low received optical power and fiber impairments.
• To demonstrate a paradigm shift from bit-level to semantic-level communication in optical fiber systems.

Proposed method

• A deep neural network is trained to extract semantic representations from source data, replacing raw bit encoding.
• Semantic symbols are directly modulated and transmitted over an optical fiber link without conventional bit-level modulation.
• The system uses end-to-end learning to optimize semantic representation and transmission jointly.
• A semantic autoencoder framework is employed to compress source data into compact semantic symbols.
• The optical link is experimentally validated under various power levels and impairment conditions.
• Performance is evaluated using semantic distortion and bit error rate metrics, comparing with conventional systems.

Experimental results

Research questions

• RQ1 Can semantic representation learning reduce information redundancy in optical fiber communication systems?
• RQ2 How does semantic transmission perform under low received optical power compared to traditional bit-based systems?
• RQ3 To what extent does semantic communication improve robustness against optical fiber impairments?
• RQ4 Can semantic symbols achieve higher spectral efficiency than conventional modulation schemes?

Key findings

• The SOFC system achieved higher information compression compared to conventional bit-based systems, reducing redundant data transmission.
• Performance was more stable in low received optical power regimes, demonstrating improved robustness.
• The system showed enhanced resilience to optical link impairments, such as chromatic dispersion and PMD.
• Semantic transmission reduced the required signal-to-noise ratio for reliable communication.
• The semantic autoencoder framework successfully preserved meaningful information with lower distortion than traditional encoding.
• Experimental results confirmed that semantic communication outperforms conventional systems in terms of spectral efficiency and error resilience.