Evolving Token Communication with Parametric Memory Network

2026-05-03

AI Summary

This work addresses the high communication overhead of existing semantic communication methods, which transmit complete semantic tokens and struggle to balance efficiency with semantic fidelity. To overcome this limitation, the authors propose an efficient semantic communication framework that transmits only token prefixes and leverages a parameterized memory network at the receiver to implicitly reconstruct the missing suffixes from the parameters of a pretrained GPT-2 model, thereby avoiding explicit memory storage. The framework incorporates a kNN-guided teacher-distribution fine-tuning module and an online evolution strategy to adapt dynamically to channel conditions and data distribution shifts. Evaluated under MIMO fading channels, the method significantly reduces communication costs and consistently outperforms state-of-the-art baselines across varying bandwidth ratios, achieving up to a 1.09 dB improvement in PSNR.

Abstract

Token communication has emerged as a promising framework for efficient wireless transmission by representing source data as compact semantic tokens. However, transmitting full semantic tokens still incurs considerable communication overhead. In this paper, we propose an evolving semantic token communication system with a parametric memory network over MIMO fading channels. Specifically, only an equal-length prefix of each semantic token is transmitted, which reduces transmission cost while preserving a consistent token structure for receiver-side recovery. At the receiver, a parametric memory network is introduced to reconstruct the missing suffix information from the received token prefixes, where semantic memory is stored implicitly in the network parameters. To realize this design, full semantic tokens are first organized into a codebook, and truncated tokens are paired with the codeword labels of their corresponding full tokens. Based on these token-label pairs, kNN-based teacher distributions are constructed to fine-tune a pretrained GPT-2-based recovery module, which learns to infer the codeword distribution of each incomplete token and recover the corresponding complete semantic token. In addition, an online evolution strategy is developed to periodically update the parametric memory network and the entire system using newly observed test samples, thereby improving adaptability under distribution shifts. Experimental results demonstrate that the proposed method consistently outperforms the existing evolving memory benchmark under different channel conditions and channel bandwidth ratios, with up to 1.09 dB PSNR improvement.

Problem

• semantic token communication
• communication overhead
• MIMO fading channels
• token reconstruction
• distribution shift

Innovation

• semantic token communication
• parametric memory network
• token prefix transmission
• online evolution
• GPT-2 fine-tuning