Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Semantic Communication for 6G Networks: A Trade-off between Distortion Criticality and Information Representability

Faizan Shafi, Rahul Jashvantbhai Pandya, Christo Kurisummoottil Thomas, Sridhar Iyer

Abstract

In this work, a self-attention based conditional generative adversarial network (SA-cGAN) framework for the sixth generation (6G) semantic communication system is proposed, explicitly designed to balance the trade-off between distortion criticality and information representability under varying channel conditions. The proposed SA-cGAN model continuously learns compact semantic representations by jointly considering semantic importance, reconstruction distortion, and channel quality, enabling adaptive selection of semantic tokens for transmission. A knowledge graph is integrated to preserve contextual relationships and enhance semantic robustness, particularly in low signal-to-noise ratio (SNR) regimes. The resulting optimization framework incorporates continuous relaxation, submodular semantic selection, and principled constraint handling, allowing efficient semantic resource allocation under bandwidth and multi-constraint conditions. Simulation results show that, although SA-cGAN achieves modest syntactic bilingual evaluation understudy scores at low SNR to approximately 0.72 at 20 dB, it significantly outperforms conventional and JSCC-based schemes in semantic metrics, with semantic similarity, semantic accuracy, and semantic completeness consistently improving above 0.90 with SNR. Additionally, the model exhibits adaptive compression behavior, aggressively reducing redundant content while preserving critical semantic information to maintain fidelity. The convergence of training loss further validates stable and efficient learning of semantic representations. Overall, the results confirm that the proposed SA-cGAN model effectively captures distortion-invariant semantic representations and dynamically adapts transmitted content based on distortion criticality and information representability for meaning-centric communication in future 6G networks.

Semantic Communication for 6G Networks: A Trade-off between Distortion Criticality and Information Representability

Abstract

In this work, a self-attention based conditional generative adversarial network (SA-cGAN) framework for the sixth generation (6G) semantic communication system is proposed, explicitly designed to balance the trade-off between distortion criticality and information representability under varying channel conditions. The proposed SA-cGAN model continuously learns compact semantic representations by jointly considering semantic importance, reconstruction distortion, and channel quality, enabling adaptive selection of semantic tokens for transmission. A knowledge graph is integrated to preserve contextual relationships and enhance semantic robustness, particularly in low signal-to-noise ratio (SNR) regimes. The resulting optimization framework incorporates continuous relaxation, submodular semantic selection, and principled constraint handling, allowing efficient semantic resource allocation under bandwidth and multi-constraint conditions. Simulation results show that, although SA-cGAN achieves modest syntactic bilingual evaluation understudy scores at low SNR to approximately 0.72 at 20 dB, it significantly outperforms conventional and JSCC-based schemes in semantic metrics, with semantic similarity, semantic accuracy, and semantic completeness consistently improving above 0.90 with SNR. Additionally, the model exhibits adaptive compression behavior, aggressively reducing redundant content while preserving critical semantic information to maintain fidelity. The convergence of training loss further validates stable and efficient learning of semantic representations. Overall, the results confirm that the proposed SA-cGAN model effectively captures distortion-invariant semantic representations and dynamically adapts transmitted content based on distortion criticality and information representability for meaning-centric communication in future 6G networks.

Paper Structure

This paper contains 22 sections, 4 theorems, 26 equations, 10 figures, 1 table, 4 algorithms.

Table of Contents

  1. Introduction
  2. Related Works
  3. Key Contributions
  4. System Model
  5. SA-cGAN Enabled 6G Semantic Communication
  6. Semantic Information Extraction
  7. KG Guided Semantic Grounding and Token Selection
  8. Channel Model
  9. Performance Metrics and Problem Formulation
  10. Performance Metrics
  11. BLEU Score (Syntactic Fidelity)
  12. Semantic Similarity (Semantic-Level Meaning Preservation)
  13. Semantic Accuracy (Token-Level Precision)
  14. Semantic Completeness (Token-Level Recall)
  15. Problem Formulation
  16. ...and 7 more sections

Key Result

Lemma 4.1

Consider the relaxed problem with $0 \le s_i \le b_i$, where $b_i = \mathbf{1}\{\alpha_i \ge \tau_\alpha\}$, and linear aggregation $\Psi(S) = \sum_{i=1}^n s_i u_i = U\mathbf{s}$, with $U = [u_1,\dots,u_n] \in \mathbb{R}^{d \times n}$, denotes the matrix of embeddings $u_i$. The relaxed optimization together with the relaxed constraints defined in Ext Opt.. If any additional surrogate terms (e.g.,

Figures (10)

  • Figure 1: System model
  • Figure 2: SA-cGAN enabled 6G semantic communication model
  • Figure 3: An illustration of source text along with the derived semantic information.
  • Figure 4: BLEU score comparison of the proposed SA-cGAN model with existing approaches
  • Figure 5: Semantic accuracy and semantic completeness versus SNR of the proposed SA–cGAN model
  • ...and 5 more figures

Theorems & Definitions (8)

  • Lemma 4.1: Convexity verification (relaxed selection)
  • proof
  • Lemma 4.2
  • proof
  • Lemma 4.3: Matroid intersection lemma (feasible augmentation under matroid intersection)
  • proof
  • Theorem 4.1: Greedy guarantee with matroid + knapsack constraints
  • proof