Integrated Computation and Communication with Fiber-optic Transmissions

2025-03-04

AI Summary

Conventional communication systems adhere to a “transmit-then-compute” paradigm, separating data transmission and computation, leading to latency and energy inefficiencies. Method: This work proposes leveraging optical fiber transmission systems natively as programmable nonlinear kernel functions, establishing an end-to-end differentiable optical-domain computing framework. Based on the nonlinear dynamical model of optical propagation, we design an implicit kernel driven jointly by dispersion and nonlinearity, enabling real-time feature mapping and computation during light propagation. Contribution/Results: To our knowledge, this is the first demonstration of directly utilizing commercial optical fiber channels as hardware-level configurable computing units, achieving physical-layer integration of communication and intelligent processing. Experiments on image classification and channel equalization show accuracy comparable to digital processors, with three orders-of-magnitude lower latency and 92% reduced power consumption—validating the feasibility and superiority of joint photonic computing–communication optimization.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Distributed Search
• Machine Learning: Hardware-aware ML

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Fiber-optic transmission systems are leveraged not only as high-speed communication channels but also as nonlinear kernel functions for machine learning computations, enabling the seamless integration of computational intelligence and communication.

Problem

• Integrate computation and communication using fiber-optic systems.
• Utilize fiber-optic transmissions for machine learning computations.
• Enable seamless integration of computational intelligence and communication.

Innovation

• Fiber-optic systems enable high-speed communication.
• Nonlinear kernel functions for machine learning.
• Seamless integration of computation and communication.