ATS-ToDMA: Adaptive Token Selection and Token-Domain Multiple Access for Cross-Modal Semantic Communications

2026-07-03

AI Summary

This work addresses the challenge of effectively managing semantic token-level multiple access and semantic interference in existing semantic communication systems. To this end, the paper proposes the ATS-ToDMA framework, which introduces— for the first time—the semantic signal-to-interference-plus-noise ratio (SSINR) as a performance metric. The framework jointly optimizes adaptive semantic token selection, token-domain multiple access, Transformer-driven interference-aware scheduling, and semantic-aware power allocation. The authors also derive theoretical bounds on semantic interference and token occupancy, along with a closed-form solution for optimal power allocation. Experimental results demonstrate that the proposed scheme significantly outperforms baseline approaches such as OMA and Semantic NOMA, achieving higher semantic throughput and decoding accuracy while simultaneously reducing total semantic interference and transmit power.

Abstract

Adaptive token processing has emerged as a promising approach for improving the efficiency of semantic communication systems. However, existing semantic communication frameworks largely overlook token-level multiple access and the impact of semantic interference among simultaneously transmitted semantic tokens. In this paper, we propose Adaptive Token Selection and Token-Domain Multiple Access (ATS-ToDMA), a novel cross-modal semantic communication framework that jointly performs semantic token selection, interference-aware scheduling, and semantic-aware power allocation. The proposed framework introduces a Semantic Signal-to-Interference-plus-Noise Ratio (SSINR) metric that captures the combined effects of channel impairments and semantic interference arising from token similarity. A transformer-based scheduler is developed to allocate selected semantic tokens across token-domain transmission slots while mitigating both intra-modal and cross-modal semantic interference. To characterize the behavior of the proposed system, analytical bounds on semantic interference and feasible token occupancy are derived, together with a closed-form approximation for semantic-aware power allocation. Simulation results demonstrate significant gains in semantic throughput and semantic decoding accuracy while reducing aggregate semantic interference and transmit power compared with OMA, Semantic NOMA, Random-TS, and Greedy ATS benchmarks.

Problem

• semantic communication
• token-level multiple access
• semantic interference
• cross-modal
• semantic tokens

Innovation

• Adaptive Token Selection
• Token-Domain Multiple Access
• Semantic Interference
• SSINR
• Cross-Modal Semantic Communication