Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Semantic Waveforms for AI-Native 6G Networks

Nour Hello, Mohamed Amine Hamoura, Francois Rivet, Emilio Calvanese Strinati

TL;DR

The paper tackles the inefficiency of bit-accurate transmission in semantic communications by introducing OSSDM, a semantic-aware waveform that encodes meaning directly in the Walsh-domain. It combines a semantic encoder (LLM+GNN) with a trainable Walsh-coefficient mapper and an end-to-end trainable waveform generator based on the Inverse Walsh Transform, enforcing spectral mask compliance through a codebook-based mapping and a joint loss that includes mutual information and semantic reconstruction terms. The approach yields an end-to-end metric, E-SSE, that captures semantic fidelity per unit bandwidth/time, and demonstrates substantial gains over conventional OFDM in both semantic robustness and spectral efficiency, especially at lower Walsh orders and under AWGN. Overall, OSSDM demonstrates the viability and advantages of meaning-aware physical signal construction for AI-native, goal-oriented 6G networks, motivating further integration of semantics, signal processing, and hardware constraints in waveform design.

Abstract

In this paper, we propose a semantic-aware waveform design framework for AI-native 6G networks that jointly optimizes physical layer resource usage and semantic communication efficiency and robustness, while explicitly accounting for the hardware constraints of RF chains. Our approach, called Orthogonal Semantic Sequency Division Multiplexing (OSSDM), introduces a parametrizable, orthogonal-base waveform design that enables controlled degradation of the wireless transmitted signal to preserve semantically significant content while minimizing resource consumption. We demonstrate that OSSDM not only reinforces semantic robustness against channel impairments but also improves semantic spectral efficiency by encoding meaningful information directly at the waveform level. Extensive numerical evaluations show that OSSDM outperforms conventional OFDM waveforms in spectral efficiency and semantic fidelity. The proposed semantic waveform co-design opens new research frontiers for AI-native, intelligent communication systems by enabling meaning-aware physical signal construction through the direct encoding of semantics at the waveform level.

Semantic Waveforms for AI-Native 6G Networks

TL;DR

The paper tackles the inefficiency of bit-accurate transmission in semantic communications by introducing OSSDM, a semantic-aware waveform that encodes meaning directly in the Walsh-domain. It combines a semantic encoder (LLM+GNN) with a trainable Walsh-coefficient mapper and an end-to-end trainable waveform generator based on the Inverse Walsh Transform, enforcing spectral mask compliance through a codebook-based mapping and a joint loss that includes mutual information and semantic reconstruction terms. The approach yields an end-to-end metric, E-SSE, that captures semantic fidelity per unit bandwidth/time, and demonstrates substantial gains over conventional OFDM in both semantic robustness and spectral efficiency, especially at lower Walsh orders and under AWGN. Overall, OSSDM demonstrates the viability and advantages of meaning-aware physical signal construction for AI-native, goal-oriented 6G networks, motivating further integration of semantics, signal processing, and hardware constraints in waveform design.

Abstract

In this paper, we propose a semantic-aware waveform design framework for AI-native 6G networks that jointly optimizes physical layer resource usage and semantic communication efficiency and robustness, while explicitly accounting for the hardware constraints of RF chains. Our approach, called Orthogonal Semantic Sequency Division Multiplexing (OSSDM), introduces a parametrizable, orthogonal-base waveform design that enables controlled degradation of the wireless transmitted signal to preserve semantically significant content while minimizing resource consumption. We demonstrate that OSSDM not only reinforces semantic robustness against channel impairments but also improves semantic spectral efficiency by encoding meaningful information directly at the waveform level. Extensive numerical evaluations show that OSSDM outperforms conventional OFDM waveforms in spectral efficiency and semantic fidelity. The proposed semantic waveform co-design opens new research frontiers for AI-native, intelligent communication systems by enabling meaning-aware physical signal construction through the direct encoding of semantics at the waveform level.
Paper Structure (5 sections, 5 equations, 4 figures)

This paper contains 5 sections, 5 equations, 4 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. Proposed Semantic Waveform Design
  4. Numerical Results
  5. Conclusions

Figures (4)

  • Figure 1: System model and key components of the semantic-aware waveform communication chain.
  • Figure 2: Performance comparison of OSSDM with OFDM-based baselines, including (i) non-semantic OFDM with LDPC channel coding and (ii) semantic symbol transmission over conventional OFDM.
  • Figure 3: E-SSE Gain of OSSDM w.r.t different OFDM numerologies, at maximal F1 score.
  • Figure 4: Semantic Waveform Spectrum w.r.t 5G Emission Mask ETSI_TS_138_104_V16_7_0.